Talk:Australia

29 percent of people in Australia Australian?

"29.06% Australian, 0.45% Australian Aboriginal, 52.8% European, 6.67% Asian, 4.69% Other, 6.33% Unstated[1]"?

Huh? How is 29 percent of the Australian population Australian?— Preceding unsigned comment added by [[User:{{{1}}}|{{{1}}}]] ([[User talk:{{{1}}}#top|talk]] • [[Special:Contributions/{{{1}}}|contribs]])

It doesn't say 29 percent of the Australian population is Australian. This section refers specifically to ethnicity. At the 2006 Census, 29% of respondents indicated that their parental ancestry was Australian, as indicated in the citation. --AussieLegend (talk) 06:54, 21 February 2009 (UTC)

I'm not sure that everyone taking the census properly understood the question at hand, then. Unless immigration of Australia is at a worldwide high, I find it hard to believe that half the population is directly European. This is where demographics becomes hard: were you born here, were your parents born here, were your grandparents born here? Sure, the caucasian population probably came from Europe at some time in the past 200 years, but when can a people start to identify themselves as being from where they are, i.e. when can these people identify themselves as 'Australian'. It seems hard to measure in this context, but maybe it's easier than I think. Maybe I misunderstood the question!Farglesword (talk) 21:56, 20 March 2009 (UTC)

The problem seems to be that some editors assume there is some scientific or rational basis for ethnic classifications. In fact ethnicity always has been based on self-identification, or arbitrary classification by others. The question asks for self-identity, which ia as valid a way to define ethnic origins as any other. --Michael Johnson (talk) 00:03, 21 March 2009 (UTC)

We have alota immigrants so australian would stand for the original settlers--Gcpeoples (talk) 05:15, 22 March 2009 (UTC)

I'd find it hard to believe that people didn't understand the question, which simply asked where their parents were born. 52.8% of responses said that one or both of their parents were born in a European country. That doesn't mean that the respondent was born in Europe so I don't see an issue with the figure. --AussieLegend (talk) 05:50, 21 March 2009 (UTC)

I think you've misunderstood the documentation here; there is no question on the 2001 or 2006 Census that asks for that level of data about where respondents' parents were born. (There's a copy of the 2006 form near the end of the online Census Dictionary, and the 2001 form is here.)

Questions 14 and 15 (13 and 14 on the 2001 form) ask "Was the person's father/mother born in Australia or overseas?" There is no option to indicate where overseas; in this data, 'country of birth of parents' is provided only at the level of 'Australia' vs 'rest of world'.

Question 18 asks "What is the person’s ancestry?" The question itself doesn't mention parents, but it points respondents at page 7 of the Census Guide for more information. That section states: "...Ancestry is not necessarily related to the place a person was born but is more the cultural group that they most closely identify with. For example, a person may be born in New Zealand but have Samoan ancestry." While it advises respondents to "consider the origins of the person's parents and grandparents for example", it certainly doesn't lock 'ancestry' in with country of parents' birth - indeed, it encourages people who are descendants of indentured South Sea Islanders to describe themselves as 'Australian South Sea Islander', even though those ancestors would have arrived in Australia about 100 years ago.

The source we're looking at, "Ancestry by Country of Birth of Parents " combines data from these three questions. However, since the analysis used to populate Australia's infobox only uses the 'total' column, it ignores all information supplied on questions 13 & 14; the only thing it represents is responses to question 18. Which doesn't "simply ask where their parents were born", not by a long shot. (Apologies for the late response to this point, but this talk page is getting pretty bloated and it's easy to miss stuff.) --GenericBob (talk) 00:48, 16 April 2009 (UTC)

And where did you get that number from, AussieLegend? I've looked at the data, and only 9, 000 Australians identified as 'European'. Most Australians of European ethnic origin identify as German, English, Italian, etc. not as just 'European'. Besides, Anglo-Celtic people that identified their ancestry as English, Scots, Irish, etc. do not form a distinct ethnic group from those who identified their ancestry as just 'Australian', and the field clearly says 'ethnic groups' not 'self-identified ancestral identity'. Aren't many European Australians of more than one European ancestry, anyway? Does the Census allow more than one answer to the "ancestry" question? Saimdusan Talk|Contribs 05:15, 1 April 2009 (UTC)

Also, where did the number for 'Asian' come from? Does it include subcontinental Indians? Does it include Middle Easterners? If it includes Middle Easterners, does it include Egyptian Arabs? If it does not, why would Lebanese Arabs, who are more related to Egyptian Arabs than to Chinese, be included in this category? Saimdusan Talk|Contribs 05:17, 1 April 2009 (UTC)

I am going to remove the ethnic groups field from the article for the time being, as how it stands now it just serves to confuse and would better be served in a section rather than just "29% Australian, 53% European, 7% Asian, etc.". It should however be added again once this issue is solved. Saimdusan Talk|Contribs 05:19, 1 April 2009 (UTC)

The data came straight from the Australian Bureau of Statistics, as indicated in the citation provided in support of the data. The ethnicities listed are summaries based on the origin of the parents, ie parents from England, Gemany, Italy, Greece etc are listed as European, Chinese, Filipino, Vietnamese are listed as Asian etc., as per convention. I really don't know where you got the rather ridiculous figure of 9,000 Europeans. There were 13,427,947 responses indicating that one or both parents were from European countries. That's a tad higher than 9,000.

There should be no confusion at all, especially if you take the time to read the citation provided. Since the information is cited and since the citation is from a reliable source I've restored it. If you can provide a better source than the ABS for the data then there might be reason to remove it, but until then it should stay. Don't delete cited information just because you didn't bother to check the citation. --AussieLegend (talk) 05:51, 1 April 2009 (UTC)

So, I did take the time to read the cited document. In the first part (2001 census) it states:

• Both parents born overseas: Responses (6,027,352) Persons (5,410,162)
• Father only born overseas: Responses (1,849,204) Persons (1,245,494)
• Mother only born overseas: Responses (1,214,965) Persons (815,324)
• Both parents born in Australia: Responses (12,258,949) Persons (9,983,897)
• Country of birth not stated: Responses (1,218,001) Persons (1,133,431)
• Total responses: Responses (22,568,471) Persons (18,588,308)

Giving the following percentages for 2001:

• Both parents born overseas: 26.71%/29.11%
• Father only born overseas: 8.19%/6.7%
• Mother only born overseas: 5.38%/4.39%
• Both parents born in Australia: 54.32%/53.71%
• Country of birth not stated: 5.4%/6.1%
• Total responses: 100%/100%

For 2006 it looks pretty similar

• Both parents born overseas: 26.51%/29.56%
• Father only born overseas: 8.32%/6.55%
• Mother only born overseas: 5.66%/4.43%
• Both parents born in Australia: 53.07%/51.79%
• Country of birth not stated: 6.44%/7.68%
• Total responses: 100%/100%

I'm not sure why the responses and persons are not the same, so I've given both numbers. It looks like the table in the article has either switched the numbers from the European and Australian parts, since its hardly possible for both parents to have been born in Australia and be European or its using the 13 Million responses you cite. This is the sum of the self-described ethnicity of the respondent, rather than where the parents were born. The citation is quite confusing though. Particularly since "Australian" isn't really a recognised ethnicity. Probably it would be better to group "Australian" with "other". AKAF (talk) 11:21, 1 April 2009 (UTC)

Firstly, there may be discrepancies between the figures in the article and what I've written above, and what I'm about to present because, once again, the ABS has changed the document linked to at the url in the citation. This was previously discussed and the article updated to reflect the more detailed data in the newer document.[1] The url is now linking to the older document again so I'll refer to the figures in that document in the following paragraphs, rather than the document that was used for the article.

The figures in each column are responses, not people. That can be confirmed simply by adding the columns and comparing the results to the two totals for each column. The only place that persons are used is in the totals section and the persons totals don't relate directly to the totals for individual ethnicities. In the right-most column (Total), there were 6,674,946 responses who identified as Australian out of a total of 22,568,471. That represents 29.57% of the total responses. The European ethnicities (Croatian, Dutch, English, French, German, Greek, Hungarian, Irish, Italian, Macedonian, Maltese, Polish, Russian, Scottish, Serbian, Spanish and Welsh) represent a total of 11,798,539, or 52.278%.

The purpose of the infobox information is to show the breakdown of the country and I think there would be more confusion if "Australian" was not identified as an ethnicity or if Australian was lumped in with "other". Either of these would give the impression that there were no people identifying themselves as Australians living in Australia and that's just plain silly. --AussieLegend (talk) 12:20, 1 April 2009 (UTC)

Yeah, the document isn't easy to read. I think the thing which people are objecting to is to referring to "Australian" as an ethnicity rather than a nationality. Certainly over half of all respondents would refer to themselves as Australian, since both parents were born in Australia. Its a bit pointless at some juncture to cite the document, when fully 1/3 of all respondents effectively have no ethnicity (~36%=Australian plus no comment). That gives an error margin of +60% on the figure for European ethnicity and +600% on the figure for Asian ethnicity. With these kinds of margins for error it would be better to leave out the data or look for a better source. AKAF (talk) 13:59, 1 April 2009 (UTC)

I think the problem here is that ethnicity is something that is difficult to define, and that self-definition is about the only reasonable way to go about it. There is no "scientific" definition of ethnic origin, and to try and impose one is to start on a slippery slope towards racism. So if people want to identify themselves in an ethnic way as "Australian" good luck to them. And no I don't think we should assume anything about what that means (ie we can't assume such people are of European decent, for instance). --Michael Johnson (talk) 23:44, 1 April 2009 (UTC)

I agree. Self identification explains why there are some peculiar figures in the source document, such as 1,455 of the 93,486 respondents who identified themselves as "Australian Aboriginal" stating that both parents were born overseas. --AussieLegend (talk) 06:00, 2 April 2009 (UTC)

The point which I was trying to make is that, although the reference exists, the quality of the data is dubious at best. I don't see that its so important that the article would fail without this piece of information. It's not like the article is struggling for length, so I would suggest removing the data from the infobox. The data needs more context, which can be better provided in the text. AKAF (talk) 06:48, 2 April 2009 (UTC)

"quality of the data is dubious at best". Do you realise that the Australian Bureau of Statistics is a Australian Government body (Similar to the United States Census Bureau) meaning that it's the most reliable source that you're going to get and find. Just because you think it's wrong that it should be removed. If you don't agree with the figures then take it up with the ABS. Bidgee (talk) 08:13, 2 April 2009 (UTC)

The data is from a very reliable source so it's not really up to us to dispute it unless we have citations from more reliable sources to the contrary. Infoboxes are supposed to summarise information from the main document. There should be discussion on it in the prose that gives it context anyway. Deleting it from the infobox just because somebody doesn't understand it is not appropriate. If we were to delete information for that reason, most infoboxes would be empty. --AussieLegend (talk) 07:02, 2 April 2009 (UTC)

A reliable source can, and does, produce unreliable data when the signal to noise ratio is high. I'm particularly commenting that its pretty brainless to quote this source as reliable for the figure of 6.67% Asian when 6.33% left their ethnicity unstated and 26% refused to comment, and left their ethnicity as "Australian". In this case the "true" figure for asian ethnicity could well be 12% or 20% and be fully consistent with the data. If we were honest in quoting the data accurately we would use error bars and divide the nonspecific data between the specific respondents, using the assumption that no specific group refused to answer this question (which assumption is invalid, but anyway). In this case the data would look something like this:

1. 0.7% $\pm$ 0.2% Australian Aboriginal
2. 81.7% $\pm$ 28.9% European
3. 10.3% $\pm$ 3.7% Asian
4. 7.3% $\pm$ 2.6% Other

The values should definitely only be quoted to one decimal place, and probably better to one significant figure. As they stand, the values are a red flashing sign that the writers of the article don't understand data analysis. Please understand that these figures are no more original research than the numbers currently in the article, since "percent ethnicity" is not a component of that document. AKAF (talk) 13:11, 2 April 2009 (UTC)

There's really not much that can be done if people don't wish to state their ethnicity. All you can do is to include the unstated figure in the list, so that the reader realises that there is a margin of error in the figures, and that has been done. I'm not sure what you mean about 26% percent refusing to comment. Only 5.6% (using the figures in the current source) refused to identify an ethnicity. As for identifying as Australian, it's already been said that there's nothing wrong with that. My parents and grand-parents were born in Australia and I identify as Australian. As far as I'm concerned that's my ethnicity. I'm sure many feel the same way. Your claims regarding percentage errors do constitue original research since they're not explicitly stated in the document and you need to make assumptions to arrive at the errors that you've presented. The figure are stated, which is why we use them. Working out percentages (not percentage errors) based on the stated figures is not original research, because no assumption is necessary. Anyone reading the document can work out that 6,674,946 is 29.58% of 22,568,471 without making a single assumption. --AussieLegend (talk) 13:54, 2 April 2009 (UTC)

The errors, sure. I mean the percentages calculated in the way above. Particularly since the percentages in the article are not quoted from the document, but are assembled from the document. AKAF (talk) 14:29, 2 April 2009 (UTC)

Of course they refused to state an ethnicity, AussieLegend. Once again, 'Australian' is not an ethnicity. White Australians who identified as Australian are not distinct from White Australians who identified as some European. I got the 9,000 figure from the data itself, the data says that 9,000 people identifed their ancestry as 'European'. Of course, it also says that there are millions of people identifying with particular European peoples. But pulling this new meaning out of the census data is WP:OR. The Aborigines are the true 'ethnic Australians' and Anglo-Celtic Australians should not perpetuate this false hierarchy with whites/aussies at the top, wogs/ethnics in the middle and abos/blacks at the bottom. I can understand seeing a second-generation paki (Punjabi) and wog (Serb, Slovene, Russian) such as myself as non-Australian but definitely Indigenous peopple are Australians! The best thing to do would be to simply show the ancestry data as it is - show the highest reported ancestries (rather than this WP:OR 'European' identity that I as a European Australian actually do support), then add a note for 'Australian'. Fixed, simple enough. Saimdusan Talk|Contribs 04:59, 3 April 2009 (UTC)

Once again, 'Australian' is not an ethnicity Why not? It seems to me to be an act of arrogance, or indeed racism, to try and appropriate people to ethnic groups they don't identify with. There is no scientific basis to ethnic definitions, only historical, cultural, political and geographical definitions, and these are both variable and numerous. So there are no hard and fast rules regarding ethnicity. If people identify as "Australian" that is what they are. And once again, no assumption can be made that they are necessarily of European decent. --Michael Johnson (talk) 23:21, 3 April 2009 (UTC)

Identifying as Australian is not refusing to state an ethnicity. It's stating that your ethnic group is Australian. The makeup of ethnic groups evolves over time. Ethnic groups aren't created one day and then stay the same for ever. People in England, many of whom descended from Roman occupiers of the British Isles aren't classified as Italian. They're classed as English. It's a similar story in Australia (and everywhere else). My family's roots in Australia date back to the First Fleet so I regard myself as Australian, not European, despite the fact that my ancestors are originally from there. Your definitions of various ethnic groups certainly seem to have some racist undertones.

As for your mythical figure of 9,000, looking at the citation I see no evidence supporting that. The smallest figure listed is 43,831 and that's for Americans.

That leaves the rather silly claim that identifying people from Europe as European is WP:OR. I'm really not sure what to say about that one. --AussieLegend (talk) 07:13, 3 April 2009 (UTC)

You had 512 relatives on the first fleet? Congratulations! Would it not be better to use the data for where the parents were born instead of the self-reported ethnicities? This is probably significantly more reliable and gives a more solid (but different) demographic picture. Just put the percentages for whom one or more parent was born in $COUNTRY. Improves the quality and still gives an idea of the immigrant makeup. For a table you don't want any data which requires multiple minutes of analysis to understand. It's far better to only report facts for which there is high-quality data. Ethnicity data will always be dubious, and no amount of massaging will help that. AKAF (talk) 07:44, 3 April 2009 (UTC)

"You had 512 relatives on the first fleet?" - What the hell are you talking about?

"Would it not be better to use the data for where the parents were born instead of the self-reported ethnicities?" - Because that isn't identified in the source data. "Overseas" is not an ethnicity. --AussieLegend (talk) 07:58, 3 April 2009 (UTC)

I was simply pointing out that you have >512 ancestors who were of reproductive age in the 1790s, to say that your ethnicity is wholly defined by one of them is the height of craziness. Many Australians have at least one Aboriginal relative (mostly this was hidden by the family) but this is not a defining trait of ethnicity, if it was more than 100 years back. The table indicates where the parents of the respondents were born. Is there another way to read it? AKAF (talk) 11:45, 3 April 2009 (UTC)

Once again, ethnicity is self defining. If people identify as Australian, it really doesn't matter if there ancestors arrived 30,000 years ago, 200 years ago, or yesterday. That's what they are. You're welcome to produce a source that argues differently. --Michael Johnson (talk) 23:21, 3 April 2009 (UTC)

Arbitrary section break

1. Even Anglo-Australian assimilationist ethnic nationalists do not see themselves as being of Australian ethnicity. They (and presumably you, AussieLegend, do too) see themselves as "white".

2. In this context, claiming that people from Europe are European IS original research. The source that is cited does not mention the word "European". Not once. If one looks at the full classification list by sex, 9, 000 people identified their ancestry as European. AussieLegend, what is your definition of European? More importantly, why should we go by your definition? Does 'European' include Turks? Why/why not? Aren't some Turks 'from' Asia? But how can one know which Turks are 'European' and which 'Asian' based on the data? Why 'European'? Why not Eurasian, or for that matter, Afro-Eurasian? Why not western/northern/eastern/southern European? 3. The cited data shows responses to the question "What is your ancestry?". The data does not mention "ethnicity". 4. Where did you get the data for 'European' anyway? Did you lump all of the 'European' respondents together? You do realize that one can select two ancestries on the census, right? This would mean that the data overlaps, so they can't be tallied together. 5. Why did you put 'Australian' respondents first? Europeans were larger. Saimdusan Talk|Contribs 06:11, 5 April 2009 (UTC)

1. "Anglo-Australian assimilationist ethnic nationalists" - Did you get that from a buzzphrase generator? I, and I'm sure most people, have no idea what that means. As for being white, what does that have to do anything?

2. I think you completely misunderstand the concept of original research, a policy that states, "Wikipedia does not publish original research or original thought. This includes unpublished facts, arguments, speculation, and ideas; and any unpublished analysis or synthesis of published material that serves to advance a position." Claiming that somebody from Europe is European is none of those things. Quite simply, since you seem to be having trouble with this, somebody from the continent of Europe is European. It's a well known, uncontroversial fact. It's somewhat ironic that you've asked what definitions I've used since there are wikilinks to the appropriate articles right next to the figures presented. I've just confirmed them for your benefit. I really don't see why you should have an issue with this. The information in the source document has been summarised into European, Asian etc because, as I've already stated, the infobox is a summary of what's in the article. Detailed lists shouldn't be included. When we did include more detail, there was an issue with people who couldn't understand that there are more Germans in Australia than Greeks now.

3. I suggest that you click on the "ANCESTRY" link in the source. If you do, you'll see that the ABS believes that "a person’s ancestry, when used in conjunction with the person’s country of birth and whether the person’s parents were born in Australia or overseas, provides a good indication of the ethnic background of first and second generation Australians". (bolding added for emphasis) Granted, the source includes everyone, but that's really not an issue.

4. You are clearly missing the point. The first and last columns are all that are relevant. The first column identifies the respondent's ethnicity, the last identifies the number of respondents who identify with that ethnicity. Where the respondent's parents were born is not important. What is important is the ethnic group with which the respondents identify.

5. Because this is an article about Australia, so Australian seems the most relevant ethnicity. The source document does the same. If this was an article about India, I'd put Indian first. It's not rocket science. --AussieLegend (talk) 07:26, 5 April 2009 (UTC)

" Anglo- Australian assimilationist ethnic nationalists " -- or ' Bogan ' Lejon (talk) 05:04, 6 April 2009 (UTC)

Yes, ancestry is an indicator of ethnicity, but that doesn't make them the same thing. The source document describes the data as 'ancestry', and in other edits you've been punctilious about sticking to the exact wording used by the ABS; let's apply the same standard here. You will notice that the Census question which provided this data refers only to 'ancestry' and doesn't refer to 'ethnicity' at any point. --GenericBob (talk) 05:30, 8 April 2009 (UTC)

Further, the percentages given in this article are not taken from the cited source; they're derived from it, apparently by dividing the number of responses in each category by 'total responses'. This is misleading, since it gives readers the impression that only 29.6% of respondents identified ancestry as 'Australian', when in fact 35.9% did so (6,674,946 of 18,588,308). --GenericBob (talk) 05:45, 8 April 2009 (UTC)

The figures are derived by simple mathematics, directly from the source data. Anyone, well almost anyone it seems, can verify the results. The figure of 6,674,946 that you've quoted is the figure for 2001, not 2006. The 2006 figure is is 7,371,824. In any case, your result is in error for 2001 because you've divided by the number of persons rather than by the number of responses. You've divided the number of apples on the bench by the number of oranges that you've eaten to determine the number of grapefruit left on the tree. That's why you've arrived at 35.9% for 2001, instead of the correct figure of 29.57%. Just to make it clear for anyone else who may look at the wrong data, the relevant cells are A56:A88 (ethnic group), G56:G88 (number of responses for the corresponding cells in column A), and G89 (Total responses). Eveything else is irrelevant. --AussieLegend (talk) 07:29, 8 April 2009 (UTC)

I was using the 2001 figures because those were the ones that you put in the article, and the percentages given were derived from those 2001 figures. Now that we've updated to 2006 figures, I'm happy to discuss those; the numbers are slightly different but the problems are unchanged.

The first is representing the data as 'ethnic groups', when the source (and the Census question on which that data is based) do not make any such representation. It's certainly related to ethnicity, but that doesn't make it the same thing. The ABS' notes on ancestry specify that it is a "good indication" of ethnic background of first and second generation Australians, when used in conjunction with the person's country of birth (which is a separate data item - Q12 on the Census form). Neither of those conditions apply here.

As an illustration of why terminology is important here, compare the responses to this question with the question on Indigenous identity. When asked "is the personal of Aboriginal or Torres Strait Islander origin?" (Q7), 425,511 people ticked 'yes' to Aboriginal ancestry (that includes 17,811 who also ticked TSI)[2]. But when asked about their ancestry (Q18 on the same form), only 115,820 people identified Aboriginal ancestry. At first glance, either of those two do look like they should be 'close enough' to ethnicity - but they disagree by half an order of magnitude! When the response is so sensitive to the way the question is worded, it's vital that Wikipedia represents that wording accurately. We should not be talking about 'ethnic group' when the source is based on 'ancestry'.

The second is that dividing by number of responses is inappropriate. When a reader sees "51.05% European, 6.33% unstated", the natural interpretation is that 51.05% of Australian people answered 'European', and 6.33% didn't answer - and this is incorrect. In fact, about 8.1% of respondents didn't answer at all. At the point where you decide that it's more appropriate to divide by responses than people, you're engaging in interpretation, and we get into the realm of original research.

I would appreciate it if you'd refrain from snideness; it's not conducive to amicable editing. And please don't remove accuracy tags until the issue has been resolved. --GenericBob (talk) 01:02, 9 April 2009 (UTC)

Dividing by the number of responses is entirely appropriate because all of the numerical data in rows 18-51 and 56-89 are numbers of responses, not numbers of people. There are only 12 cells (B52:G52 and B90:G90) that refer to persons. There is no explanation in the source of how the figures in rows 52 and 90 were determined, only an explanation of why they do not correspond the the numbers of responses (cells A97 and A98). If you tally the figures in each column you will see that the totals correspond to the values in rows 51 and 89, not 52 and 90. You CAN NOT divide the number of responses by the number of persons, as you did to arrive at the incorrect figure of 8.1% instead of the correct figure of 6.33% because, as is stated in the source, the sourcee "is a multi-response table, and therefore the total responses count will not necessarily equal the total persons count." Dividing the number of responses by the total number of persons is not just OR, it's WP:SYNTH. --AussieLegend (talk) 08:40, 9 April 2009 (UTC)

Can you explain why all other computations on the data are WP:SYNTH while your computation is not? AKAF (talk) 09:08, 9 April 2009 (UTC)

More precisely: Could you please provide the row and column number where in the source it is explicitly stated that 51.05% of Australians are European, as required by WP:SYNTH? AKAF (talk) 09:11, 9 April 2009 (UTC)

WP:Primary sources states: All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source, rather than original analysis of the primary-source material by Wikipedia editors. It then names census results as an example. In the preface to the ASC:CEG

It is difficult to define ethnicity in a way that is both useful and generally acceptable and it is not the function of this document to attempt an extensive definition of the concept. However, because the words ‘ethnicity’ and ‘ethnic’ are associated with many different meanings it is useful to provide some definitional material. The Macquarie Dictionary (Third Edition, 1997) provides the following meanings:

1. relating to or peculiar to a population, especially to a speech group, loosely also to a race. 2. relating to the origin, classification, characteristics, etc., of such groups. 3. of or relating to members of the community who are migrants or descendants of migrants and whose native language is not English. 4. recognisable as coming from an identifiable culture.
For the purposes of ASCCEG it is sufficient, and not controversial, to say that the term ‘ethnicity’ refers to the shared identity or similarity of a group of people on the basis of one or more factors, including the following which were enunciated in a report entitled The Measurement of Ethnicity in the Australian Census of Population and Housing [Borrie report, 1986]

... Since the publication of the Borrie Report in April 1984, the multicultural nature of Australian society has further developed resulting in a more sophisticated and enlightened approach to cultural diversity. Nevertheless, ... Dennis Trewin ASC:CEG, 2000-01

It is appropriate to challenge any information interpreted or extrapolated from data in a primary source, even if it was given some context in the article. The data was collated and massaged, apparently, "to satisfy wide community interest in the ethnic and cultural composition of the Australian population and the characteristics of particular migrant community groups.", and it is tempting synthesise that interest with Australia's historical (or hysterical) xenophobia and obsession with race. One need look no further than earlier Australian censuses to see 'arbitrary sectarian breakdowns', to put a twist on this section's heading. Inclusion of the breakdown in the article would be highly questionable, requiring secondary sources, it is unacceptable to thrust it upon a reader as infobox content. The data presented in these garish oversimplifications is, at best, next to useless as encyclopaedic information. The 'ethnic' section of this infobox, in a featured article, is effectively operating as a POV and OR fork and should be removed.

It should be obvious to anyone that dividng one type of figure (responses) by another (persons) to come up with a third type (percentage) is completely wrong when the source document says that the two don't correlate. In order to determine the percentages you need to be using the same data, which I am. As I already explained above, "There is no explanation in the source of how the figures in rows 52 and 90 were determined, only an explanation of why they do not correspond the the numbers of responses". The figures for persons have come from somewehere else, even though they are included in the source used. It is not possible to arrive at the figures presented for persons using the data contained in the source. This is not the case when determining percentages based on responses. Again, as I've already stated above, "The figures are derived by simple mathematics, directly from the source data. Anyone, well almost anyone it seems, can verify the results". WP:SYNTH does not apply because only one source is being used for the calculations, which you can confirm yourself, unless you have absolutely no idea where Europe and Asia are. If that's the case you can always follow the links that are provided. --AussieLegend (talk) 09:43, 9 April 2009 (UTC)

AussieLegend, it might seem sensible at first to divide responses by total responses. However, it is complete nonsense to put percentages derived this way in the infobox. These percentages are the percentages of responses in each category, but a normal reading of the infobox would suggest that it is giving percentages of people identifying with each ethnicity. It is not possible to do this, since while the ABS has made sure people are only counted once in each of their categories, there are sure to be quite a few counted twice in the supercategories listed in the infobox. Summarising data should not be covered by WP:SYNTH, but any suggestion that this sort of summary of these response figures give percentages of various ethnicities is simply wrong. JPD (talk) 11:38, 9 April 2009 (UTC)

If the ABS has counted each person once in each of the categories, and the sum of the categories in which each person is counted once is the total of each supercategory, how can each person be counted more than once in the supercategory but not in the individual categories? That's like saying 1+1+1+1+1=6. I know what you're getting at but the error that may exist is so low as to be inconsequential. The figures are really a guide anyway, since they are based on data that is now 2.5 years old. Census data is only accurate on the day of the census. I can't see why people are getting their knickers in such a twist. According to Australia's population clock, we have 1,871,101 more people in the country, a 9.4% increase, and we have no idea what ethnicity they are. There's bound to be a big dent in the figures that anyone here could calculate. As I said, it's just a guide. --AussieLegend (talk) 12:35, 9 April 2009 (UTC)

Yes, supercategories are a problem - there is no way to tell how many distinct people identified ancestry somewhere in Europe. The data you are using is not suited for the use to which you're trying to put it, and in attempting to pound a round peg into a square hole you're engaging in OR. --GenericBob (talk) 14:49, 9 April 2009 (UTC)

You're under a misapprehension. The fact that there isn't a one-to-one match between 'responses' and 'persons' does not mean we can't meaningfully divide them. Let's try a simpler example: I ask Bob and Jane to pick up to two words that describe their cars. Bob picks "red" and "Ford"; Jane picks "red" and "Holden". Two people, four responses (two "red", one "Ford", one "Holden"). So how would we represent this in the Bob and Jane's Cars Infobox?

Under your approach - dividing responses by total responses - we would represent this as "50% red, 25% Ford, 25% Holden". Under mine, it would be "100% red, 50% Holden, 50% Ford" - which is a rather more accurate representation of the situation. There is no problem with dividing a count of responses by the count of persons when what you're looking for is the proportion of persons who gave a particular response.

And yes, my approach most certainly is OR. I believe it's a higher quality of OR - I have a fair bit of professional familiarity with this sort of work - but it certainly doesn't belong in article space. Nor does yours. The mathematical side of the analysis is OR, because of this issue of interpretation, and conflating 'ancestry' with 'ethnic group' is also OR. --GenericBob (talk) 11:58, 9 April 2009 (UTC)

Under your approach the breakdowns total 127.98%. Put that in the infobox and see what reaction you get.--AussieLegend (talk) 12:35, 9 April 2009 (UTC)

This is only a problem if you believe these categories are mutually exclusive - in which case, you're in trouble, because the question used to collect them is specifically designed to allow for a person falling into more than one category. In any case, as I've already said above, I have no intention of putting them in the infobox. You seem to have decided that this infobox MUST be filled, and that one way or another the ABS 'ancestry' data is an appropriate source with which to achieve that; both of these are bad assumptions. --GenericBob (talk) 14:49, 9 April 2009 (UTC)

There's no analysis of the data involved. It's just a summary of the presented data in an easier to read format than:

• • Australian - 7,371,824
  • Australian Aboriginal - 115,280
  • American - 56,283
  • Chinese - 669,890
  • Croatian - 118,046
  • Dutch - 310,089
  • English - 6,283,650
  • Filipino - 160,374
  • French - 98,333
  • German - 811,540
  • Greek - 365,147
  • Hungarian - 67,628
  • Indian - 234,722
  • Irish - 1,803,741
  • Italian - 852,418
  • Lebanese - 181,745
  • Macedonian - 83,983
  • Maltese - 153,803
  • Maori - 92,912
  • New Zealander - 160,681
  • Polish - 163,802
  • Russian - 67,054
  • Scottish - 1,501,201
  • Serbian - 95,365
  • Sinhalese - 73,856
  • South African - 79,521
  • Spanish - 84,322
  • Turkish - 59,393
  • Vietnamese - 173,658
  • Welsh - 113,250
  • Other(c) - 1,397,647
  • Ancestry not stated - 1,609,443
  • Total Responses(d) - 25,410,601

which doesn't fit in an infobox too well. --AussieLegend (talk) 09:51, 9 April 2009 (UTC)

No, it doesn't. Neither would the explanation of how they arrived at those figures. 10:10, 9 April 2009 (UTC)

• In order to prove your claim that you have performed no synthesis from the data, could you please provide the row and column number where in the source it is explicitly stated that 51.05% of Australians are European. This is not visible in the excerpt which you quote above. AKAF (talk) 11:21, 9 April 2009 (UTC)

I agree that trying to fit all that into an infobox would be a horrendous mess. The correct solution is not to violate WP:OR, but to leave that field empty until such time as an appropriate source can be found. --GenericBob (talk) 11:58, 9 April 2009 (UTC)

After checking the ABS site in search of a more appropriate source I believe we actually have one already. WP:OR has not been violated. There should be no problem. --AussieLegend (talk) 12:41, 9 April 2009 (UTC)

Please provide the row and column number where in the source it is explicitly stated that 51.05% of Australians are European. This is not visible in the excerpt which you quote above. Doing this will demonstrate that your synthesis is not WP:OR. AKAF (talk) 10:06, 10 April 2009 (UTC)

How many times does this question have to be answered before you stop acting like a broken record? The data presented is simply a summary from a single, cited source, directly converted to a percentage for ease of comparison so WP:OR doesn't apply. --AussieLegend (talk) 11:45, 10 April 2009 (UTC)

No, it's not 'directly' converted to a percentage - before making that conversion you've converted the data into supercategories e.g. 'European'. Given the multi-response nature of the data, the assumption that it is valid to sum individual categories to get a meaningful supercategory is not a trivial one, which makes it OR. Choosing between dividing by number of responses or number of persons is an analytical decision, and that makes it OR. Interpreting 'ancestry' as equivalent to 'ethnic group' - also OR.

Since this is chewing up a lot of space on the Talk page and it doesn't look like we're getting any closer to agreement, I'm going to put this up for RFC. --GenericBob (talk) 13:47, 10 April 2009 (UTC)

Stating thst it's "not 'directly' converted" is being extremely pedantic. The numbers are just being added based on continental location for ease of display. --AussieLegend (talk) 13:57, 10 April 2009 (UTC)

Second arbitrary section break

Since you haven't answered any of my questions, I'll just ask them again.

AussieLegend, what is your definition of European? More importantly, why should we go by your definition? Does 'European' include Turks? Why/why not? Aren't some Turks 'from' Asia? But how can one know which Turks are 'European' and which 'Asian' based on the data? Why 'European'? Why not Eurasian, or for that matter, Afro-Eurasian? Why not western/northern/eastern/southern European? 3. The cited data shows responses to the question "What is your ancestry?". The data does not mention "ethnicity". 4. Where did you get the data for 'European' anyway? Did you lump all of the 'European' respondents together? You do realize that one can select two ancestries on the census, right? This would mean that the data overlaps, so they can't be tallied together.

1. "Anglo-Australian assimilationist ethnic nationalists" - Did you get that from a buzzphrase generator?

Anglo - Of or relating to England or Britain. Australian - Of or relating to Australia. Therefore, an Anglo-Australian is an Australian of British/English ancestry. Assimilationist - a proponent of assimilation. Ethnic nationalism - an ideology where the nation is defined by ethnicity.

As for being white, what does that have to do anything?

Oh come on, "whiteness" is one of the main factors of European ethnic identity.

I think you completely misunderstand the concept of original research, a policy that states, "Wikipedia does not publish original research or original thought. This includes unpublished facts, arguments, speculation, and ideas; and any unpublished analysis or synthesis of published material that serves to advance a position." Claiming that somebody from Europe is European is none of those things. Quite simply, since you seem to be having trouble with this, somebody from the continent of Europe is European. It's a well known, uncontroversial fact. It's somewhat ironic that you've asked what definitions I've used since there are wikilinks to the appropriate articles right next to the figures presented. I've just confirmed them for your benefit. I really don't see why you should have an issue with this.

Did you include Turks or not? Don't sidestep the question. It IS original research to claim that 50% of Australians are part of a European ethnic group, as no source claims that. These people are not 'from Europe', many of the people who identified with a European ancestry on the census have Australian ancestors going back to the early 1900s or even to the 1800s (I know this for a fact, as I know many of these people).

The information in the source document has been summarised into European, Asian etc because, as I've already stated, the infobox is a summary of what's in the article. Detailed lists shouldn't be included. When we did include more detail, there was an issue with people who couldn't understand that there are more Germans in Australia than Greeks now.

Well, the way the data is presented now is even more confusing except to Anglo-Celtic supremacists.

I suggest that you click on the "ANCESTRY" link in the source. If you do, you'll see that the ABS believes that "a person’s ancestry, when used in conjunction with the person’s country of birth and whether the person’s parents were born in Australia or overseas, provides a good indication of the ethnic background of first and second generation Australians".

Giving a strong indication of their ethnic background does not DETERMINE their ethnic background.

4. You are clearly missing the point. The first and last columns are all that are relevant. The first column identifies the respondent's ethnicity, the last identifies the number of respondents who identify with that ethnicity. Where the respondent's parents were born is not important. What is important is the ethnic group with which the respondents identify.

Yes, and none of them identify as European (well, 9,000 do). You obviously just lumped overlapping categories together.

::5. Because this is an article about Australia, so Australian seems the most relevant ethnicity. The source document does the same. If this was an article about India, I'd put Indian first. It's not rocket science.

Not many Indians would identify as ethnic Indians on their census. Of course, the Indian census only has caste and tribal identification but if it did have ethnic identification "Indian" would be as rare as those Yugoslavs who identified their nationality (the term used in Yugoslavia for ethnicity at the time) as"Yugoslav". So, maybe 2% of the population. Anyway, it has to be either in order of immigration to Australia (in which case Aborigines would be first) or in order of size (in which case Europeans would be first) otherwise it's just confusing (and you seem to dislike a previous table that confused people who were unaware of the fact that there was major German immigration here in the early 1900s).Saimdusan Talk|Contribs 10:48, 10 April 2009 (UTC)

"Since you haven't answered any of my questions" - Clearly, you haven't been reading my responses because I have answered your questions.

"what is your definition of European" - responded to here, 32 minutes after you asked.

"Does 'European' include Turks? Why/why not? Aren't some Turks 'from' Asia?" (etc) - There was really no need to respond to this since you answered your own question. Obviously, since it's not possible to determine which Turks were from Asia and which were from Europe, they've naturally been included with "others". You could have answered this with a calculator.

"Why not Eurasian, or for that matter, Afro-Eurasian? Why not western/northern/eastern/southern European?" - Also responded to here, as well as several other places before and after you asked the question. The information provided is a s-u-m-m-a-r-y. If I was to include terms like Eurasian, Afro-Eurasian or Anglo-Australian assimilationist ethnic nationalists, I might as well have left the data as was presented, which is not appropriate for an infobox. The continental boundaries provide a far more simple method of grouping people which is appropriate for an infobox where everything is summarised.

"The cited data shows responses to the question "What is your ancestry?". The data does not mention "ethnicity"." - responded to here, 75 minutes after you asked the question.

"Where did you get the data for 'European' anyway?" - See response to "what is your definition of European"

"You do realize that one can select two ancestries on the census, right?" - responded to here, 75 minutes after you asked the question.

Responses to new statements & questions

"Anglo - Of or relating to England or Britain." (etc) - So you created the term yourself.

"Oh come on, "whiteness" is one of the main factors of European ethnic identity." - . That's a rather racist attitude to take. Surely the main, and really the only factor, is location, hence the name.

"Did you include Turks or not?" - This is a question that you shouldn't need to ask. If I had included 56,000 turks the percentages would have been different, something easily verified with a basic calculator.

"It IS original research to claim that 50% of Australians are part of a European ethnic group, as no source claims that." - The cited source document provided exact figures for each of the ethnicities chosen by respondents. The percentage is a summary of data extracted straight from that data. It's not OR.

"many of the people who identified with a European ancestry on the census have Australian ancestors going back to the early 1900s or even to the 1800s " -

"I know this for a fact, as I know many of these people" - Now that is classic WP:OR. It's specifically mentioned in the policy, under "Citing oneself". How many of these almost 13 million people do you know and can you provide a citation from a reliable source to confirm your claim. Don't sidestep this. I want proof.

"Well, the way the data is presented now is even more confusing except to Anglo-Celtic supremacists." - More racist statements. It'd be quite ridiculous if it wasn't for the racist tone.

"Giving a strong indication of their ethnic background does not DETERMINE their ethnic background." - Why not?

"Yes, and none of them identify as European (well, 9,000 do)." - You keep stating this figure of 9,000 but so far you've failed to cite the source other than by vague statements such as "I've looked at the data" and "I got the 9,000 figure from the data itself". When are you going to provide a source for this mythical data?

"You obviously just lumped overlapping categories together" - No, I didn't obviously do that. Each identified ethnicity has been included only once.

"Not many Indians would identify as ethnic Indians on their census." - The fact is, you don't know how anyone other than you identified.

"...as rare as those Yugoslavs who identified their nationality" - Hardly relevant as Yugoslav isn't mentioned in the source data.

"Anyway, it has to be either in order of immigration to Australia " - Why? This is an article about Australia so it makes more sense to first identify those who consider themselves Australian since that is what the other ethnicities will be compared against in the first place.

"you seem to dislike a previous table that confused people who were unaware of the fact that there was major German immigration here in the early 1900s" - Perhaps you should do a little more research before commenting. What actually happened was that somebody changed the order of ethnicities.[3] Somebody else too exception to that.[4] The someone else changed Greek to German.[5] I thought that was strange, considering that Greek has always been a significant ethnicity in Australia, but, not being one to act on gut instinct, unlike some,[6] I decided to do some research. Finding the existing citation woefully inaccurate and out of date, found a better reference and updated the article accordingly,[7], listing the top 5 ethnic groups. Over subsequent weeks, "Greek" kept being added back in.[8] I reverted the first one, suggesting we keep it to the top 5 groups,[9] but Greek kept being re-added. I eventually asked myself if it was OK for the top 5 European ethnicities, why isn't it OK for the 6th, 7th etc and why shouldn't the same be the case for Asian, or any ethnicities. Unfortunately, this would blow out the size of the infobox, which is undesirable so I decided the best and fairest option would be to remove the top 5 European ethnicities altogether. Guess what? It worked. It had nothing to do with Germans and skin colour certainly didn't come into it. Sorry to disappoint you. --AussieLegend (talk) 13:54, 10 April 2009 (UTC)

Oh yes, a simple calculation of percentages. Now I believe you. AKAF (talk) 14:42, 10 April 2009 (UTC)

As you should. (a+b+c+d+e+f+g+h+i+j+k)/l isn't that hard to do, even on a calculator. A primary school kid should have no problems. --AussieLegend (talk) 14:50, 10 April 2009 (UTC)

Fine, then perhaps you will reference your source for considering Russians to be European, but Turks not to be. You appear to resist all efforts to explain that what you have done is distill a complex synthesis based on data of dubious value into a single number. Look, original research is a problem, because it defeats the purpose of correct referencing. You need to quote only exactly what your secondary sources say. If you can find a secondary source which analysis your primary source to say what you want then you're golden. Just quote the secondary source. As it is, you are being the secondary source by quoting something which is not exactly stated in the primary source. You can see this just by noting how many ways there are of calculating that percentage in the thread above. You need to just deal with the fact that the source doesn't say what you want and move on. AKAF (talk) 18:11, 10 April 2009 (UTC)

"perhaps you will reference your source for considering Russians to be European" - I'm happy to concede that Russians being included in European is an error. Thanks for picking it up. I've now fixed the error. It made a 0.26% error, although if I'd stuck to single digit accuracy as was suggested it would only be 0.2%. Funny how nobody picked it up considering I posted that almost 10 days ago.

"You need to quote only exactly what your secondary sources say." - If it was in the prose of an article I'd almost agree with you but we can't be exact in the infobox because it takes up far too much space, which is why a summary in an easy to read format is used. The reason I'd only almost agree with you is that we can't quote exactly what a source says. We're forced not to in order to avoid copyright violations. The summary is simply the same data presented in a different format. You need to deal with that. --AussieLegend (talk) 19:30, 10 April 2009 (UTC)

"Funny how nobody picked it up considering I posted that almost 10 days ago" - not really. Wikipedia editors are not here to check the details of your analysis - if it's complex enough to have a significant risk of error (which apparently it is) then it shouldn't be here at all.

In a lot of cases, copyright would require us to paraphrase sources, but this isn't one of them, because ABS content is published under the CC-BY license. So this is a non-issue (see WP:CFAQ).

The fact that the source can't be fit into the infobox without a great deal of shoehorning isn't an argument for doing that shoehorning, it's a sign that the source is not an appropriate way to fill the infobox. --GenericBob (talk) 00:38, 11 April 2009 (UTC)

"Wikipedia editors are not here to check the details of your analysis " - That's not a convincing argument because that's exactly what people, including you, have been doing, at least until it comes to responding to what I've written here.

"if it's complex enough to have a significant risk of error (which apparently it is)" - No it isn't apparent because the error was exceptionally minor (0.2%).

"The fact that the source can't be fit into the infobox without a great deal of shoehorning isn't an argument for doing that shoehorning, it's a sign that the source is not an appropriate way to fill the infobox" - Again, not a convincing argument. As you should well know, and not only because I've mentioned it several times, infoboxes are only summaries. There's very little source information that fits in an infobox without being summarised down to only a few words.

Yup, I've heard you say it repeatedly - and if it was an accurate summary that stuck to things that could be verified from the source, I wouldn't object. It's not, and lack of space is not a good enough reason for an inaccurate summary. --GenericBob (talk) 14:26, 14 April 2009 (UTC)

Note also that the presentation of the data indicates a precision of 0.01%; in that context, an error of 0.2% is not minor. FWIW, this false precision is also a problem with the analysis. The ABS makes it clear that there are all sorts of reasons why the counts published are unlikely to be exact - e.g. nonresponse, problems in coding answers, etc etc - and it's generally up to the secondary source to estimate how badly those impact on derived data. This sort of issue is one of the reasons why WP:RS warns about primary sources. --GenericBob (talk) 22:18, 15 April 2009 (UTC)

So you created the term yourself.

The term "Anglo"? Surely you've heard this term before!

Why not?

Read your own bloody source. It does not say "it determines their ethnicity". It says it gives a "GOOD INDICATION".

The fact is, you don't know how anyone other than you identified.

Um... I'm talking about the Indian census at the moment. Indians would definitely identify as Indians in the Australian census. In India however,each group is heavily divided. No, I'm not suggesting to put this into the article so it doesn't need a citation. I don't think it makes sense for you to deny such the obvious fact of India's diversity.

Hardly relevant as Yugoslav isn't mentioned in the source data.

This time, I was talking about the Yugoslav census. Remember,the article about India? I'm saying that people that identified their ethnicity as "Indian" would be a small minority, and definitely shouldn't go on the top (whatever the plurality group is, probably Bengalis as Bengali is the largest Indian language other than Hindi). You're the one that used the example of the India page, so I don't see how it's "hardly relevant".

I'm happy to concede that Russians being included in European is an error.

What about Greeks? There are Greeks native to Asia, you know. Ever heard of Greek Cypriots and Pontic Greeks?

That's a rather racist attitude to take. Surely the main, and really the only factor, is location, hence the name.

How is that racist? It's like saying that Afro-diasporic people don't have a concept of 'blackness', or that Islam isn't the largest religion of the Punjabi people. It's simply an observation about the IDENTITY of 'indigenous' Europeans and the European diaspora when they (well, we) distinguish ourselves from groups that tend to be darker-skinned and European mongrels such as myself.

One of the most important things with your data is that the numbers for 'Europeans' and 'Asians' CAN'T be added together as hundreds of thousands of Australians identified with more than one ancestry on the census.

I'm still having trouble seeing where you got your definitions from, either. If you define Asian as 'people from Asia', then we're fine, but this is about ethnic groups. Turks and Lebanese are definitely not seen as 'ethnic Asians' in Australia, as with the US 'Asian' refers to East and Southeast Asians. Similarly, Turks and Lebanese are generally not seen as 'Whites' (essentially meaning fair-skinned Europeans in this country, which is why it IS a factor of European ethnic identity) either. In fact, Greeks and Italians are only 'sometimes' seen as Whites, often being distinguished from the 'Aussies' (really, Anglo-Celtic people) as being 'Wogs'. None of this is mentioned in your source, nor does it say 'Asian', 'European' or anything else even ONCE. Saimdusan Talk|Contribs 00:22, 11 April 2009 (UTC)

"The term "Anglo"? Surely you've heard this term before!" - Of course I've heard of anglo before. As you well know, the term at issue was "Anglo-Australian assimilationist ethnic nationalists". You're just being silly with responses like that.

"It says it gives a "GOOD INDICATION"." - So, what's wrong with that? The people who responded identified the ethnic group they most identify with.

"Um... I'm talking about the Indian census at the moment." - Um... Why? The Indian census has absolutely no relevance to this issue.

"This time, I was talking about the Yugoslav census." - Again, why? How is this relevant to this or the Indian article?

"You're the one that used the example of the India page, so I don't see how it's "hardly relevant"." - I raised the Indian page only in response to your question regarding why Australian was at the top of the list. OK let's forget India. Blow it off the map. Use New Zealand instead. (Finally, a use for New Zealand!)

"What about Greeks? There are Greeks native to Asia, you know. Ever heard of Greek Cypriots and Pontic Greeks?" - We can't make assumptions that this or that response refers to Greek Cypriots, Pontic Greeks, Babylonian Italians or Martian New Zealanders. If a response says Greek then it's in the Greek box and Greece is a European country. To make any assumptions otherwise is original research. Look at the responses for "Australian Aboriginal". 1,455 responses claim that bothe parents were born overseas. How many Aboriginals have both parents born overseas? To assume that these people aren't really Aboriginal is original research. We have to include them in the Australian Aboriginal group whether we agree with it or not.

"How is that racist? It's like saying that Afro-diasporic people don't have a concept of 'blackness'," - They might have a concept of blackness but blackness isn't an an indicator that they're Afro-disaporic any more than whiteness is an indicator that somebody is European. There are plenty of Europeans of all colours.

"One of the most important things with your data is that the numbers for 'Europeans' and 'Asians' CAN'T be added together as hundreds of thousands of Australians identified with more than one ancestry on the census." - Again, this goes back to the Aboriginal example. If people identify with a particular group in the source data then we have to include them in that group or else you're resorting to original research.

"If you define Asian as 'people from Asia', then we're fine," - We're fine then.

" Turks and Lebanese are definitely not seen as 'ethnic Asians' in Australia," - As already indicated, Turks are included in others. So are Lebanes. It seems we agree.

"Similarly, Turks and Lebanese are generally not seen as 'Whites' (essentially meaning fair-skinned Europeans in this country, which is why it IS a factor of European ethnic identity)" - Again, I have to disagree with your fixation on skin colour. People from the US are generally white and they're not European. --AussieLegend (talk) 07:15, 14 April 2009 (UTC)

Australia a Continent???

This aritcle states that Australia is a continent. Forgive me for saying this, but there are only 7 continents in the world which consist of Asia, Africa, Europe, North America, South America, Antarctica, and Oceania. Therefore, Australia is not a continent and stating such in this article is false. Australia is nation within Oceania. --Yoganate79 (talk) 02:40, 8 March 2009 (UTC)

Please see the first sentence of the Continent article and its supporting citation. Nick-D (talk) 02:44, 8 March 2009 (UTC)

As well as the numerous discussions about this very issue in the Talk page archives. Oceania is not a continent. It is a region. --AussieLegend (talk) 06:47, 8 March 2009 (UTC)

Australia is the continent in the list you give, not Oceania, most of which is small island nations. --Dmol (talk) 08:01, 8 March 2009 (UTC)

Having been involved in s similar discussion on this page, I'd like to suggest to Yoganate that he leaves it alone. "Continent" is a poorly defined term, which you can pretty much let mean what you want. So just let those editors continent themselves with their inaccuracy. Australians have been taught in school that "Australia is the largest island and the smallest continent", and pretty much everyone who edits this page believes that to be an important part of the red white and blue running through their veins. Leave alone that a goodly amount of what gets taught in school is bullshit: "Australia is the only place where marsupials are found" or "Australia was one of the first countries to have universal suffrage" occur immediately as examples. It appears to bother these editors not one whit that this argument leaves Tasmania, New Guinea, and most of Indonesia unattached to any continent. Moreover it appears to bother noone that this entire set of statements predates continental plate theory, Or that this statement was made as the protectorate of New Guinea was a part of Australia. One might as well write the whole article from the point of view of 1960s historians, with the noble attempts at integration of the poor abo kids into generous white families. However seeing as the statement about continents is absolutely unbased in any reality, and its just a throwaway line, its probably not worth it to argue over such a triviality. Except that its location in the first paragraph it gives the reader the impression that the scholarship for the whole article will be as slipshod. AKAF (talk) 15:10, 15 March 2009 (UTC)

I remember you. You were the editor who claimed that the Australian Government's Department of Foreign Affairs and Trade website was an advertsing website and not a reliable source.[10] Welcome back and please remember, should you choose to present any arguments, rather than just attack the editors of this article and Australians in general, some civility will result in your comments gaining a better response than the incivility that you've just demonstrated. --AussieLegend (talk) 16:06, 15 March 2009 (UTC)

Here's the truth about the government: they have departments whose purpose is to promote some aspect of Australia to the country and overseas (ie "Advertising", for some value of that word). And yes, the department of foreign affairs and trade is not a reliable reference for the same reason that an autobiography is not the authoritative guide to a person's life. In fact most government departments are not good references for an article about that department's government, and should be regarded as a primary source for the purposes of citing. So you'll excuse me if I find referencing such a site to be at best disingenuous. AKAF (talk) 08:18, 16 March 2009 (UTC)

Continent discussions - for future reference

Past "Is Australia a continent?" and "Australia is not a continent" discussions, and related topics, may be found at:

• Talk:Australia/Archive 3#In Australasia or only country in a continent
• Talk:Australia/Archive 3#Excessive categories
• Talk:Australia/Archive 4#The whole continent?
• Talk:Australia/Archive 5#Australia or its commonwealth
• Talk:Australia/Archive 6#Island
• Talk:Australia/Archive 8#Landmass
• Talk:Australia/Archive 13#Australia is not a continent
• Talk:Australia/Archive 14#Australia is the only continent occupied by a single country
• Talk:Australia/Archive 14#Is Australia a country or a continent?
• Talk:Australia/Archive 14#Australian Continent

I could have sworn it's been discussed more than that. --AussieLegend (talk) 09:15, 8 March 2009 (UTC)

It really is quite sad that Australians are brainwashed and uneducated to think that their country is one of the classical continents found on earth. My question is, if Australia is considered a continent, then what continent does New Zealand belong to then? In any case, Greenland should also be classified as a continent then too. Let's classify Long Island, Cuba, Japan, and the British Isles their own continents too. --Yoganate79 (talk) 01:52, 26 March 2009 (UTC)

Excuse me? Brainwashed and uneducated? How insulting! Bidgee (talk) 01:57, 26 March 2009 (UTC)

Exactly australia has better education systems than america. At least we can go to school without getting shot.

Australia a continent yes but australia is the figurehead of australasia a.k.a. oceania new zealand, papua new guinea and that fall under that continent.— Preceding unsigned comment added by 124.171.106.175 (talk • contribs)

Australasia is not "a.k.a." Oceania, it's a region of Oceania. Neither are continents. --AussieLegend (talk) 09:16, 9 April 2009 (UTC)

Countries do not have to belong to a continent. Sorting out which continent the Maldives is located would be tricky. Furthermore, Australia is not a third world country with a tyrannical government. It terms of public brainwashing and lack of education, I'd encourage you to visit someday, and see if your comment is justified. ∗ \ / (⁂) 11:46, 9 April 2009 (UTC)

remains a Commonwealth realm

Do Editors think it is appropriate to say in the opening para that Australia "remains a Commonwealth realm"? Shouldn't the article say Australia is a "constitutional monarchy" - as that term is used in the UK article and Canada's. This is being discussed on Talk:United Kingdom. Regards. Redking7 (talk) 11:48, 15 March 2009 (UTC)

I support using the "Constitutional monarchy" description in the opening para:

1. Redking7 (talk) 11:48, 15 March 2009 (UTC)
2. Hawkeye7 (talk) 03:25, 18 March 2009 (UTC)
3. Slac speak up! 04:52, 25 March 2009 (UTC)
4. --Michael Johnson (talk) 07:30, 30 March 2009 (UTC)
5. Night w (talk) 06:13, 22 April 2009 (UTC) (P.S. I also don't think it's a big deal).
6. [Insert your username here]

I support using the Commonwealth realm" description in the opening para

1. [Insert your username here]

Comments/Questions I need more detail in why should it be changed, whats more widely used (Commonwealth realm or Constitutional monarchy) in Australia (not just within the Commonwealth). Bidgee (talk) 11:52, 15 March 2009 (UTC)

I'd also like to know why it should be changed. I should point out that Wikipedia works on consensus, not voting, so the polls here and at Talk:United Kingdom are premature. It's not appropriate to make a post on one article's talk page and declare consensus at another, which is what prompted Bidgee's reversion.[11] You need to actually discuss the issue before consensus can be reached and this particular matter needs to be discussed here. --AussieLegend (talk) 12:26, 15 March 2009 (UTC)

Have a look at Talk:United Kingdom - its being discussed there too (more editors are involved on that page). Regards. Redking7 (talk) 21:59, 16 March 2009 (UTC)

The discussion at Talk:United Kingdom is about use at United Kingdom, not here. As I indicated on your talk page,[12] any consensus achieved at Talk:United Kingdom does not immediately become applicable here. You need to discuss the issue here, not there and explain why the change should occur here. --AussieLegend (talk) 01:19, 17 March 2009 (UTC)

I support the reference. However, the purpose of it is to get across that when the colonies federated in 1901, the new country did not "go it alone" as a republic, but continued to be part of the Commonwealth (or Empire as it was then) and continued to acknowledge the British monarch as its monarch. It's also a constitutional monarchy of course, but maybe the purpose of this reference to Commonwealth realms could be spelled out a little more explicitly. -- JackofOz (talk) 23:52, 16 March 2009 (UTC)

Constitutional monarchy is a widely understood term. Commonwealth realm is a term not used in Australia, is not widely understood, and is not as descriptive as constitutional monarchy. Hawkeye7 (talk) 03:30, 18 March 2009 (UTC)

Yes, but just saying it remained/became a constitutional monarchy could mean that it set up its own monarchy, distinct from the British monarchy. We need to communicate the continuity of Australia's allegiance to the British monarchy between pre-1901 and post-1901. -- JackofOz (talk) 07:51, 18 March 2009 (UTC)

What do the sources say? I'm and Australian, and I can't say that I've ever seen a description of Australia as being a 'Commonwealth realm'. A Google search of 'Australia Commonwealth realm' doesn't produce any reliable-looking sources in the first couple of pages of results ([13]), and the Queen's website states that "Australia is a constitutional monarchy with The Queen as Sovereign": [14], though Australia is in the drop down list of 'Commonwealth realms' at the top of the page. Nick-D (talk) 07:59, 18 March 2009 (UTC)

Well I've found this (Paper crown when Charles takes reign) which seems to point that Australia is a Commonwealth realm but again I request for more detailed information with sources before I can make a decision. Bidgee (talk) 08:15, 18 March 2009 (UTC)

Why can't both have a place in the lead? Here is a suggestion:

“ For around 40,000 years before European settlement commenced in the late 18th century, the Australian mainland and Tasmania were inhabited by around 250 individual nations of indigenous Australians. After sporadic visits by fishermen from the immediate north, and European discovery by Dutch explorers in 1606, the eastern half of Australia was claimed by the British in 1770 and initially settled through penal transportation to the colony of New South Wales, founded on 26 January 1788. The population grew steadily in the following years; the continent was explored, and during the 19th century another five largely self-governing Crown Colonies were established. On 1 January 1901, the six colonies became a federation, and the Commonwealth of Australia was formed.

The Commonweath is a constitutional monarchy with strong liberal democratic traditions. As a Commonwealth realm, Australia's monarch is Elizabeth II. The population is just over 21.7 million, with approximately 60% concentrated in and around the mainland state capitals of Sydney, Melbourne, Brisbane, Perth, and Adelaide. The nation's capital city is Canberra, located in the Australian Capital Territory (ACT).

”

It is just one possibility. -Rrius (talk) 03:29, 26 March 2009 (UTC)

I agree, I don't see why it should be a case of one or the other. I think your suggestion is an excellent way of incorporating both truths. --AussieLegend (talk) 03:44, 26 March 2009 (UTC)

For the information of RedKing, "Commonwealth Realm" is a *very* infrequently used term in Australia, even in constitutional law circles (my constitutional law textbook didn't mention it once). "Constitutional monarchy" is vastly more common; I would even go so far to say that this is an imported Canadism, like "Her Majesty's Loyal Opposition" (never used in any official context in Australia). Slac speak up! 02:27, 30 March 2009 (UTC)

We are, nevertheless, both things. We're a constitutional monarchy because we're governed by a constitution and we have a monarch (regardless of who that monarch is); but because the monarch happens to be Elizabeth II, we're also a Commonwealth realm. Not all constitutional monarchies are Commonwealth realms, but all Commonwealth realms are constitutional monarchies. -- JackofOz (talk) 05:13, 30 March 2009 (UTC)

Well Australia is many things, but we don't have to include them all in the opening paragraph. Commonwealth realm may be an accurate description, but it is not one that strikes me as being commonly used in relation to Australia. I'm not fussed, but I can't see the point of including it. --Michael Johnson (talk) 07:30, 30 March 2009 (UTC)

It may not be particularly often used in Australia, but if we write the article using only terms that are frequently used in Australia, it would be of less value to international readers. There seems little point in having an article called Commonwealth realm without mentioning - not just in that article but also in this one - that Australia is such an animal. -- JackofOz (talk) 12:06, 30 March 2009 (UTC)

Can't it just say ...maintained a stable liberal democracy and a constitutional monarchy within the Commonwealth. (or something similar and better phrased)? Why not use both terms and get rid of that vague "political system" term? Night w (talk) 06:20, 22 April 2009 (UTC)

citation needed

The article is currently a FA, but how come it still has [citation needed] templates in a few sections? Aaroncrick^{(Tassie Boy talk)} 09:04, 16 March 2009 (UTC)

Because it is a high visibility article that is edited by a hell of a lot of people, making it difficult to maintain at FA standard for any length of time. This was featured in June 2005. It probably doesn't deserve to be featured now. Hesperian 01:22, 17 March 2009 (UTC)

The FA criteria probably wasn't as strict back then. Although I must admit I was a bit surprised when i saw that is was a FA. Aaroncrick^{(Tassie Boy talk)} 09:41, 18 March 2009 (UTC)

What about removing the info? That normally sparks a reaction - hopefully to find the cite. --Merbabu (talk) 23:01, 20 March 2009 (UTC)

Well there's a thought! Never thought about it that way. Aaroncrick^{(Tassie Boy talk)} 09:56, 7 April 2009 (UTC)

Category "Liberal democracies"

I think there is no more absolute consensus that Australia is a liberal democracy beacuse of strict Internet censorship. Therefore I will remove this article from category Category:Liberal democracies. —ilaiho (talk) 17:53, 22 March 2009 (UTC)

User:AussieLegend undid my edit, saying there is only one censored website. According to some recent news and that Wikipedia article, I have understood that there is much more irresponsible censorship than "one censored website". Anyway, this categorization is not very important for me. —ilaiho (talk) 17:07, 24 March 2009 (UTC)

The United States, United Kingdom and New Zealand all censor the internet, and remain on the category Category:Liberal democracies. Hawkeye7 (talk) 22:05, 24 March 2009 (UTC)

The idea that the government's censorship regime disqualifies Australia as a liberal democracy is absurd (as much as I may detest its planned internet filter scheme). Plenty of liberal democracies have more stringent censorship than Scandinavian countries (in fact, most do). Slac speak up! 04:54, 25 March 2009 (UTC)

Firstly, the ludicrous internet filtering proposal is currently a trial and is not (yet) an actual reality. Secondly, as the article on Liberal democracy points out, all such political structures do contain specific limits on specific freedoms. Every liberal democracy contains some internet restrictions on (for example) child pornography. As the article on the topic makes clear, the core of a liberal democracy is not any particular social law but the citizens elective rights, including free and fair elections and a competitive political system. Euryalus (talk) 09:47, 7 April 2009 (UTC)

How can a country with the Queen as a state head be a democracy??? And liberal??? Sorry but it doesn't work together. Monarchy is not a democracy. —Preceding unsigned comment added by 61.69.27.220 (talk) 06:22, 14 April 2009 (UTC)

Demography

Under the demography section, the information that only 70 aboriginal languages still exist is incorrect. According to the 2005 Ethnologue, which is the most authoritative linguistics survey, there are 280 recorded aboriginal languages, 238 still spoken and 42 extinct. I don't know how to change this since the article is protected. Also, there is no source for the claim that an indigenous language is the main language for 50,000 people. Gordon, Raymond G., Jr. (ed.), 2005. Ethnologue: Languages of the World, Fifteenth edition. Dallas, Tex.: SIL International. Online version: http://www.ethnologue.com/. Claritarejoice (talk) 17:55, 23 March 2009 (UTC) Claritarejoice

Australia Interactive Map

I suggest the Link below containing Google Interactive Map of Australia . http://www.all-maps.info/2009/03/australia-map.html —Preceding unsigned comment added by Kewaga (talk • contribs) 14:20, 30 March 2009 (UTC)

Business Address?

Can you please add Australia's business address to this article. It is 1601 MASSACHUSETTS AVE NW C/O AUSTRALIAN EMBASSY WASHINGTON DC 20036 and Australis's company number is 0000805157. Source - http://www.sec.gov/cgi-bin/browse-idea?CIK=0000805157&action=getcompany

Umm, no, I don't think so. --Michael Johnson (talk) 01:30, 3 April 2009 (UTC)

Please keep in mind that Australia is also well known as the ' Down Under ' and the ' Wide Brown ' . Lejon (talk) 04:49, 6 April 2009 (UTC)

Never heard of Wide Brown. Aaroncrick^{(Tassie Boy talk)} 09:57, 7 April 2009 (UTC)

What are they teaching kids at school these days?

I love a sunburnt country,

A land of sweeping plains,

Of ragged mountain ranges,

Of drought and flooding rains.

I love her far horizons,

I love her jewel-sea,

Her beauty and her terror

The wide brown land for me!

If you say you've never heard of that..... --AussieLegend (talk) 10:31, 7 April 2009 (UTC)

I think it's a good idea, but maybe we should post it to America's business address. Um, what's the US' business address? The US business address is being moved to Shanghai, right?--Merbabu (talk) 10:35, 7 April 2009 (UTC)

Nah, it's somewhere in Calcutta. -Rrius (talk) 12:34, 7 April 2009 (UTC)

Australia?Austria?

Maybe we should put a

Not to be confused with Austria.

here.People often make such mistake.Especially non-English native user like me.囧--半弯不直男 (talk) 13:54, 8 April 2009 (UTC)

RfC: Is data in infobox 'ethnic groups' field adequately supported by source?

A user has requested mediation on this issue. A mediator will be here shortly to assist you. The case page for this mediation is located here.

Template:RFCsoc

Summary: Original source[15] presents results of a Census question in which respondents were allowed to give up to two responses to the question "What is the person's ancestry?" Source gave (a) total number of responses for each of several fine-level categories (e.g. "Scottish", "Italian", and "Ancestry not stated"), (b) total responses (approx. 25 million), and (c) total persons (approx. 20 million - discrepancy due to people providing more than one response.)

Infobox on Australia uses this data in the 'ethnic groups' field. Categories have been converted to broader supercategories not explicitly named in the source (e.g. "European", "Asian") and data are expressed as percentages rather than totals. Percentages are calculated as the ratio of responses in relevant categories/supercategories to total responses.

Some points of contention are:

- Is it original research to represent data compiled and published on the basis of 'ancestry' as 'ethnic groups'?

- Is it OR to categorise individual countries into broader entities? (e.g. deciding whether to count Turkey as part of 'Europe'?)

- Is generating totals for supercategories by summing fine-level categories a routine calculation, or does the complexity of the data (specifically, the potential for a single person to be counted in more than one fine-level category due to multiple responses) make this OR?

- Is converting totals to percentages a routine calculation, or does the complexity of the data (specifically, the question of whether to divide by total responses or total persons) make this OR?— Preceding unsigned comment added by GenericBob (talk • contribs)

I suggest that you read the "Quality Statement - Ancestry" section of this ABS article (linked to from within the source document), which states "The ancestry data collected in the Census is a measure of self-identification of ethnic or cultural group affiliation and therefore provides a broad measure of cultural diversity."" There's the link to ethnicity.

Another interesting statement is "Scottish ancestry (in terms of responses for both ANC1P and ANC2P) has risen from 2.4% of persons in 2001 to 5.9% of persons in 2006"." It's interesting because if you divide the number of responses reporting Scottish ancestry by the number of total responses (534,882/22,568,471 and 1,501,201/25,410,601 for 2001 and 2006 respectively) you arrive at the results of 2.4% and 5.9%, the figures quoted in the ABS document. On the other hand, if you divide by total persons, as GenericBob would have us do, the percentages arrived at are 2.9% and 7.6%, which are significantly different to the figures quoted by the ABS. The method used to arrive at the figures in the infobox is the same used by the ABS, i.e. responses divided by total responses. --AussieLegend (talk) 15:16, 10 April 2009 (UTC)

Regarding the issue of counting Turkey as part of Europe, this was not done so I don't see why it is a point of contention. As I have indicated, and as can be confirmed by converting the percentages back to raw data, Turkey has been included in "Others" as the Turkish responses can not be reasonably included in either Europe or Asia without making assumptions. --AussieLegend (talk) 15:24, 10 April 2009 (UTC)

It is important for wikipedians to understand proper academic citation and the value of excluding material which is nice for the narrative developed for the reader, but which is mightily truthy. If you reference something from a source, you can't analyse the data from the source to get your result. You can look at a secondary source which does that analysis, but wikipedia is not about original analysis, or WP:OR in the local jargon. It is pretty obvious from the comment above that analysis is happening. I find it unlikely that any of the participants on this page are qualified demographers, and if so, it wouldn't matter, since there are other places for demographers to publish their original research. AKAF (talk) 18:25, 10 April 2009 (UTC)

You are being ridiculous in claiming that analysis is being done. The only "analysis" being done is to group ethnicities by continent, which requires absolutely no analysis. Is England in Europe? Is Germany in Europe? Is Japan in Asia? Is Vietnam in Asia? These are the only questions being asked and the answers require no analysis. --AussieLegend (talk) 19:03, 10 April 2009 (UTC)

So, is Russia in Europe? The fact that you had to reconsider this issue should be a pretty strong clue that this is nontrivial analysis.

I agree that the quality statement certainly supports the idea that ethnic group affiliation is related to ancestry data, I believe it's OR to carry this to the point of treating them as the same thing - note that "ethnic or cultural group" in there. As I noted earlier, the ABS documentation says that ancestry is a "good indication" of ethnic background of first and second generation Australians, when used in conjunction with the person's country of birth.

Do you believe the data we're discussing here pertains only to first- and second-generation Australians? Do you believe that it's being used in conjunction with the person's country of birth? (Hint: country of birth is a separate data item, collected in a different question, and none of the data from that question is included in the source document you're citing.)

Scots issue - I'm going to delay my answer to that until after the public holiday, since it will require giving somebody a phone callcontacting someone during business hours. --GenericBob (talk) 23:37, 10 April 2009 (UTC)

OK, I've discussed this issue with a staff member in the ABS' Census Products & Services section, noting that the verbal description of the Scots figures ("x% of persons") is inconsistent with the data (apparently calculated as a percentage of responses). Her response was that "it would be far more sensible to calculate based on persons rather than responses", and after checking with those responsible for the quality statement, it will be amended from this:

"Scottish ancestry (in terms of responses for both ANC1P and ANC2P) has risen from 2.4% of persons in 2001 to 5.9% of persons in 2006 and this may be partly due to its inclusion in the list of check box responses..."

to this:

"Scottish ancestry (in terms of responses for both ANC1P and ANC2P) has risen from 2.9% of persons (2.4% of all responses) in 2001 to 7.6% of persons (5.9% of all responses) in 2006. This may be partly due to its inclusion in the list of check box responses..."

Not sure how long it will take to get a corrected version on the website, but in the meantime, if you're not content to take an anonymous Wikipedia editor's word for it ;-) you can verify this by contacting the ABS. Alternately, while I'd rather not give out my real name here, I'd be happy to provide a copy of the email exchange to one of the Aussie admins. Edit 20/4/09: This correction has now appeared on the ABS website, worded as above.

Note that this quality statement was never presented as a statement on how percentages should be calculated - the purpose of this section was to explain a different issue (i.e. the fact that adding a checkbox for 'Scottish' had greatly increased the number of such responses - something that's apparent regardless of whether you divide by total responses or persons). This is probably how the error managed to slip through, and it's a reminder of why we should stick to things that are explicitly stated in the source rather than attempting to extrapolate from sources that weren't written - or checked - with that particular extrapolation in mind. --GenericBob (talk) 05:13, 14 April 2009 (UTC)

Note - since it makes more sense to respond to the following paragraph by paragraph, I've broken up AussieLegend's comments and added his signature to each paragraph to keep attribution clear. If this is a problem, please feel free to restore the original structure and move my responses to a single section at the bottom. --GenericBob (talk) 07:33, 14 April 2009 (UTC)

"The fact that you had to reconsider this issue" - I didn't reconsider anything. A mistake was identified and I immediately corrected it because I had obviously just placed Russia in the wrong pile. All the calculations were done on a spreadsheet so it's an easy thing to do. I note that you never challenged it so, using your logic, you must have agreed that Russia is in Europe.

"Do you believe the data we're discussing here pertains only to first- and second-generation Australians?" - The source doesn't just identify 1st and 2nd generation Australians. It deals with every Australian.

"if you're not content to take an anonymous Wikipedia editor's word for it" - Not just me, it's policy. We're not reliable sources so what you or I say means nothing unless we can back it up with a citation from a reliable source. I've cited a reliable source showing that I used the same methodology as the ABS to calculate the percentages.

"Note that this quality statement was never presented as a statement on how percentages should be calculated " - However, it does identify the methodology used by the ABS which is extremely relevant. It was clear from the Census dictionary statement that the ABS determined percentages by dividing responses by responses. In order to summarise the data supplied by the ABS it's necessary to use the same methodology in order to come up with the same figures. To not do so, eg dividing responses by persons, is original research. If the ABS do change the census dictionary then the methodology will have to change because then, and only then, will dividing responses by responses become OR. Until that time, divividing by responses is appropriate, dividing by persons is OR. --AussieLegend (talk) 05:54, 14 April 2009 (UTC)

OK, the ABS have now corrected the page, making the changes that I indicated last week - are we all now agreed that choosing 'total responses' as the appropriate measure is OR? --GenericBob (talk) 04:41, 20 April 2009 (UTC)

Supercategories

AL: "The fact that you had to reconsider this issue" - I didn't reconsider anything. A mistake was identified and I immediately corrected it because I had obviously just placed Russia in the wrong pile. All the calculations were done on a spreadsheet so it's an easy thing to do. I note that you never challenged it so, using your logic, you must have agreed that Russia is in Europe.

Er, no. I make no statement either way on whether Russian (or Turkish) ancestry counts as 'European'. Those are knotty questions, and I believe any such assessment would be OR; whether Russians and Turks are 'European' depends on the individuals and on definitions used. Silence does not imply agreement, it only implies that I have a finite amount of time to spend here. Checking the details of your calculations - when my point is that those calculations don't belong here in the first place - would not be a good use of my time. --GenericBob (talk) 07:33, 14 April 2009 (UTC)

Equation of 'ancestry' and 'ethnic group'

AL: "Do you believe the data we're discussing here pertains only to first- and second-generation Australians?" - The source doesn't just identify 1st and 2nd generation Australians. It deals with every Australian.

Agreed. So do you acknowledge that we can't justify equating 'ancestry' to 'ethnic group' by relying on a source which only draws a connection between the two in the context of first and second generation Australians and when used in conjunction with the person's country of birth? --GenericBob (talk) 07:33, 14 April 2009 (UTC)

Amendment to source material

AL:"if you're not content to take an anonymous Wikipedia editor's word for it" - Not just me, it's policy. We're not reliable sources so what you or I say means nothing unless we can back it up with a citation from a reliable source. I've cited a reliable source showing that I used the same methodology as the ABS to calculate the percentages.

You have misunderstood WP:RS - the sourcing standards given there are a necessary criterion for inclusion, not a sufficient one. It states that content included in Wikipedia should be supported by 'reliable sources', as defined therein; it does not state that everything found in such sources should be used. (That would be an 'infallible source', and at that point we're arguing religion, not demography.)

In this case, you have cited a 'reliable source' that (when taken along with the included figures) contradicts itself. The wording currently says that the figure given is a percentage "of persons", while we can deduce from the figures that it was calculated as a percentage of total responses. The fact that it contradicts itself makes it pretty clear that one of those two things is in error - either 'persons' was the correct method and the calculation was mistakenly done on total responses, or 'total responses' was the correct method and the documentation is using the wrong terminology.

It should be abundantly clear that a source that contradicts itself on a specific issue is not reliable on that particular issue, no matter how well it satisfies the letter of WP:RS. That on its own, without any clarification from ABS, is a pretty good reason not to use it for Wikipedia content. The fact that several other editors have argued against your approach should be another good reason to realise that what you're doing is nontrivial interpretation - and hence OR. If you're determined to keep arguing the point until that clarification comes out, you'll have a third good reason - but the longer you keep digging, the longer it takes to climb out afterwards. --GenericBob (talk) 07:33, 14 April 2009 (UTC)

Scope of source material

AL:"Note that this quality statement was never presented as a statement on how percentages should be calculated " - However, it does identify the methodology used by the ABS which is extremely relevant. It was clear from the Census dictionary statement that the ABS determined percentages by dividing responses by responses. In order to summarise the data supplied by the ABS it's necessary to use the same methodology in order to come up with the same figures. To not do so, eg dividing responses by persons, is original research. If the ABS do change the census dictionary then the methodology will have to change because then, and only then, will dividing responses by responses become OR. Until that time, divividing by responses is appropriate, dividing by persons is OR.

It was also clear from the Census dictionary statement that the ABS felt it was appropriate to discuss percentages 'of persons' (and I think we are agreed that, whatever they might be, the percentages currently on Australia are not percentages of persons?) I have said this several times already, but apparently I need to repeat it once more: I am not arguing for the infobox to include figures calculated by dividing by persons - even when the ABS quality statement is corrected, that would still be OR because of the other issues mentioned above (namely, equation of 'ancestry' to 'ethnic group' and formation of supercategories). I am arguing that the ABS source we're discussing cannot be used to fill this data field without committing OR. The only reason I'm mentioning alternate (and better) ways to analyse that data is to underline the fact that the sort of demography you're doing here is not as straightforward as you believe. --GenericBob (talk) 07:33, 14 April 2009 (UTC)

I'd like to take a moment and point out that it's considered poor form to respond the way you have,[16] despite the explanation. I'm assured that responding as I did here, quoting the protions of the respective comments that you wish to reply to, is preferred. Editing others comments is frown upon in general[17] and, even though you have given permission, it really shouldn't be up to me to re-edit your comments so that they're in the appropriate form. --AussieLegend (talk) 07:56, 14 April 2009 (UTC)

It's not universally verboten; for instance, as mentioned in the last of your links there, "In some cases, it is OK to interrupt a long contribution, either with a short comment (as a reply to a minor point) or with a heading (if the contribution introduces a new topic)." I felt this was a long contribution that could usefully be divided into subtopics without taking you out of context (and IMHO this discussion is painfully long already), but per your request I've re-edited to restore your original comment in full.

By the same token, perhaps you could refrain from snide edit summaries[18][19][20] in future? --GenericBob (talk) 08:50, 14 April 2009 (UTC)

(offtopic, but...) GenericBob, often people do replies as you were thinking about above, by quoting parts of the text of the other person in italics. This gets around the problem of editing other people's comments. I personally find AussieLegend's method of quoting to be confusing and counterproductive because he takes too little of the original text and so I end up spending half my time trying to decide who wrote what. I personally prefer to either quote a larger portion, or to quote nothing at all. Personal preferences vary greatly across the project. AKAF (talk) 09:16, 14 April 2009 (UTC)

You don't need to quiote everything the person has said, just the important points. Quoting everything just makes the discussion ridiculously long, as does creating numerous Level 2 headings for each point of a reply. --AussieLegend (talk) 09:38, 14 April 2009 (UTC)

Yep, I set the header levels wrong. Should be fixed now. --GenericBob (talk) 14:20, 14 April 2009 (UTC)

I never disagreed with that. I just noted that I find your style of quoting hard to read.AKAF (talk) 10:26, 14 April 2009 (UTC)

The point is, you didn't divide into sub-topics. You just broke apart a response by one editor. You haven't even divided into sub-topics now. You've broken out into individual main topics which makes it even more confusing to follow. As for the snide comments:

[18] - The same editor continued to ask the same question multiple times, even though answers to the question had been given multiple times. The edit summary wasn't snide, it was a summary of what I'd expressed in the edit.

[19] - Not snide, it was an attempt at some humour, based on what the editor had said and the way he had presented himself. You really need to be far less serious and try to laugh occasionally.

[20] - It was you who tried to introduce a different methodology for calculating percentages to that used by the ABS, creating confusion and trying to introduce original research while criticising me for doing the same thing, when all I'd done was summarised presented figures from the source using the same methodology used by the author of the source. If you're going to be critical of somebody, be critical of yourself, both for that and for not acknowledging, after I showed it to be the case, that your methodology was not the same. The edit summary expresses exactly what had happened. Your maths was bad. --AussieLegend (talk) 09:38, 14 April 2009 (UTC)

"trying to introduce original research"... nope. The only content change I have ever advocated to this section is removing the material currently in this part of the infobox. I have certainly argued that there is a better approach than the one you took, but as I've explained several times already - I'm sure other editors are getting tired of hearing me say this - I am not arguing that that better approach should be used to populate the infobox. I mention it to illustrate that this is not a trivial issue, and in deciding to go with one half of a self-contradicting source rather than the other half you are performing OR.

I have never contested the claim that the "2.4%" figure was derived by the 'total responses' method - while there is nothing on the ABS website (currently) to confirm it, I'd guess you're correct in concluding that that was the method used. What I do contest is your assumption that an undocumented method (contradicted by the documentation, at that) that could only be determined by reverse-engineering a calculation used only to discuss the effects of a particular checkbox on the form should be treated as the ABS' prescription for how figures should be calculated in another context. Doesn't that seem even a teensy weensy bit like OR?

Sometimes even the ABS website has errors. In this case, you've been unfortunate enough to latch onto the error first instead of the correct approach described in the same page - which is the sort of thing that happens to all of us once in a while. But when other editors point out that there are numerous problems with your method - and that the ABS website contradicts itself on this issue - and that you can very easily call the ABS to confirm that you're using the wrong method - trying to save face by accusing them of incompetence is an unwise tactic. As it happens, I work with this sort of stuff for a living (and I'd be happy to substantiate that to any admin willing to drop me an email address); while there are plenty of people around who can teach me things about demography, I've seen enough here to satisfy me that you're not one of them. --GenericBob (talk) 14:17, 14 April 2009 (UTC)

(unindent) Actually you still haven't answered the question I asked multiple times, so I feel that your snook is unreasonably cocked. In a tertiary source, you can't go re-analysing the data from a primary source, so you have only two possibilities:

1. Either: WP:OR is violated.
2. Or: You can show where the source says exactly what you quote.

What we "source" in wikipedia are directly attributable data, which can be traced back to verbatim quoted from the source material. So show me the verbatim quote or remove the material as WP:OR. There is no third choice. AKAF (talk) 10:35, 14 April 2009 (UTC)

• I made a change to the infobox here. cygnis insignis 10:54, 14 April 2009 (UTC)

Since we seem to be at an impasse, I'm sticking up a request for mediation. Further context on this dispute is in Talk:Australia#29_percent_of_people_in_Australia_Australian.3F above and Talk:Australia#Removal_of_Ethnic_Group below.

Removal of Ethnic Group

None of the groups in the "ethnic group" category are actually ethnic groups. Most of them are geographic designations...and the term "Australian" is a heritage term at best.

Intranetusa (talk) 04:41, 16 April 2009 (UTC)

You're going to need to do a lot better than that to justify removal. As has been explained above, the information in the infobox is a summary of ethnicities based on geographic origin. Summaries are used because information in the infobox needs to be brief. The CIA Fact Book, which is widely accepted as a reliable source, identifies Australian, Asians and Europeans in its list of ethnic groups.[21] Notable, reliable sources such as Encyclopædia Britannica interchange the terms "ancestry" and "ethnicity", even within the same document.[22] --AussieLegend (talk) 09:14, 16 April 2009 (UTC)

Removed again since the data is unsourced (see about half of this talk page). AKAF (talk) 09:35, 16 April 2009 (UTC)

The data is sourced. There's a citation attached, the methodology for summarising the data has been explained and proven and the question about ethnicities has been supported with references to two very reliable sources. There has been no consensus reached to remove this information. Ethic information has been in the article for a long lime so, until there is consensus to remove it then it shouldn't be removed again. --AussieLegend (talk) 10:28, 16 April 2009 (UTC)

But the citation is bullshit, as has been explained to you by a large number of people. However I'm at the point where you win this WP:GAME. Best of luck with it. AKAF (talk) 10:38, 16 April 2009 (UTC)

"The CIA Fact Book, which is widely accepted as a reliable source, identifies Australian, Asians and Europeans in its list of ethnic groups." No, it does NOT. The groups it lists are "white 92%, Asian 7%, aboriginal and other 1%". (Note that those percentages also conflict with the ones currently in the article that you're trying to defend - you've listed "5.17% Asian".)

"the methodology for summarising the data has been explained and proven" - assertion is not proof. I've repeatedly pointed out the fallacy of using 'ancestry' alone as 'ethnic group' when the source makes it clear that ancestry should be considered in combination with other information, and you've repeatedly ducked the issue. I've repeatedly pointed out that a source that contradicts itself on an issue (as the ABS page currently does on 'persons' vs 'total responses' method) cannot be considered reliable on that issue, and you've repeatedly ducked the issue. And so on and so on.

"until there is consensus to remove it then it shouldn't be removed again" - you might want to read WP:CCC, which specifically explains why this argument doesn't hold water. In any case, you seem to be the only person still arguing for inclusion of this material - and 'consensus' is not 'unanimity'.--GenericBob (talk) 12:17, 16 April 2009 (UTC)

"No, it does NOT" - Yes, it does SO. Look in the Ethnic Groups column for Norfolk Island where it says "descendants of the Bounty mutineers, Australian, New Zealander, Polynesian". You need to look at everything and not be so limited and selective in your focus. I'll let you find European on that page. I did. --AussieLegend (talk) 12:39, 16 April 2009 (UTC)

Yes, but to quote directly from that source [23] "Ethnic groups: white 92%, Asian 7%, aboriginal and other 1%". See, that's the difference between synthesis and proper citation, namely that I can directly quote the source, and it says what I want. Now do that for your source. Is there actually more than one person who believes that the current version is correct? AKAF (talk) 16:50, 16 April 2009 (UTC)

And based on the same page you've cited, the July 2009 (yes, 3 months from now) population estimate is 21,262,641. That's great except that, as of right now, the official Australian population clock[24] estimates the population to be 21,733,346, almost 471,000 more than the CIA's estimate. Based on the current net increase of 1 person every 90 seconds, which is what it has been for some time now, we should have around 21.8 million on 1 July, 542,000 more than the CIA estimate. Maybe the CIA knows something the ABS doesn't. Yes, that page may say what you want, but what it says is wrong. --AussieLegend (talk) 17:49, 16 April 2009 (UTC)

Ding, ding, ding. Yes Alex I know the answer! What is a WP:PRIMARY source? AKAF (talk) 18:35, 16 April 2009 (UTC)

Are you really going to argue that the CIA has a better idea of Australia's population than the ABS does? --AussieLegend (talk) 23:26, 16 April 2009 (UTC)

A bit more detail on population estimation, and I apologise for the length of this digression:

First off, there are several different ways to define the 'population' of Australia. Do you count foreign diplomatic staff? US military personnel passing through? Australian residents who are currently overseas? The answer depends on what you want to do with the data (or, sometimes, on what you can most consistently collect.)

Every five years, the ABS runs a census. This counts most of the people in Australia, with a few deliberate exclusions (e.g. foreign diplomatic staff, IIRC). A substantial number of people get missed for one reason or another, and a few people get counted two or even three times; the net result of this is that the Census undercounts people.

To counter this, shortly after the Census, the ABS runs a post-enumeration survey - they sample people, match them to Census responses, and use this to estimate things like overcount and undercount (for instance, when the PES finds people who weren't recorded on Census), as well as various other quality-control purposes.

Census doesn't record people who were overseas at the time, but we're interested in them, so the ABS estimates the number of Australian nationals who are out of the country, and adds them towards the population.

The result of this is that a year or two after the Census was run, the ABS has a pretty good idea of the resident population on Census night (most recent being mid-06).

What if you want to know the population for June 30, 2007? To get a high-quality estimate ('estimated resident population', or ERP), you want to look at net births, deaths, and migration since then. Ideally you would just count registrations of births, death, and migration, and adjust by the net total - but in practice, not all of this data is recorded in a timely fashion, so a bit more analysis is required, involving various assumptions and corrections. (This can be quite tricky - for instance, because people needed to register births to qualify for the 'baby bonus', there was a spike in birth registrations at that time.)

This process takes time, so again you can only get these estimates in hindsight, and they're adjusted as more data comes in - the estimates for June 2007 won't be finalised until some time around 2012 or 2013, when the data from the 2011 Census has been processed.

If you want to estimate the population in the future, you have to start with a recent good estimate (e.g. the latest ERP), and make some assumptions from there about what births, deaths, and migration might look like in the future. Since the ABS staff aren't magicians, they can't do that with anything approaching 100% accuracy, so they produce a large range of possible values for these variables. (As the ABS disclaimer on this data notes, "The projections are not predictions or forecasts, but are simply illustrations of the growth and change in population which would occur if certain assumptions about future levels of fertility, mortality, internal migration and overseas migration were to prevail over the projection period.") Combining possible choices for those assumptions gives a large number of scenarios (currently 72); of these, the ABS picks three based on mid-range, high-range, and low-range assumptions. In the ABS' words, "These are intended to illustrate a range of possible future outcomes, although there can be no certainty that any particular outcome will be realised, or that future outcomes will necessarily fall within these ranges."

If you want to estimate the population right now... you have to do the same thing, because data on today's births/deaths/migration won't be available for quite some time yet. That's how the 'population clock' figures work, using the assumptions made for Series B.

The result of all this is that the 'population clock' figures - like any other projections - are not guaranteed to be particularly accurate. To be blunt, the population clock is there to give the casual public something they can look at without having to read through a lengthy description of what we really know (or don't) about the Australian population. If you want high-quality figures for the past, you need to use the ERP figures. If you want high-quality figures for the present, or the future, you need to wait until it becomes the past... and then go to the ERP figures. (And if you need to make decisions about the future now, you'd better be good at dealing with data where even the degree of uncertainty is uncertain.)

When you combine that uncertainty with the question of how exactly you choose the scope for a population, a discrepancy of around 550k is not particularly remarkable. Maybe the CIA picked a different set of assumptions for projection purposes. Or their update schedule means they're working off an earlier ERP figure. Or the 'population' they're counting excludes some category of people who are included in the ABS projections (e.g. Australians who don't live in Australia, or non-citizens who do.) It doesn't mean the Factbook staff are incompetent, just that they made some different choices when dealing with difficult issues requiring a lot of professional judgement.

What would be a mistake is to attempt to combine the two sources, using one as a guide to how the other should be interpreted, because there's no guarantee that their working definitions are consistent with one another's, and neither source was designed to be compatible with the other. --GenericBob (talk) 00:31, 17 April 2009 (UTC)

To be honest, I didn't read all of the above because there's just too much waffle there. I did skim it and it's obvious that you've missed the point, again. The ABS is the agency that collects all of the census data and issues the reports on the data collection. Those reports are used by every government and most non-government agencies, as well as international agencies. You won't find international agencies doing their own census. They use the data, and usually that means the ABS reports, for information about Australia. One of the tools that the ABS provides is the population clock which is reasonably accurate, despite it just being a projection. You can see this when new data from a census is released. The ABS explains how it arrives at the population data, the CIA doesn't. My point was that the CIA projection is so significantly (a lot more than 0.2%) different to the official figures that it lacks credibility. --AussieLegend (talk) 00:52, 17 April 2009 (UTC)

"the population clock which is reasonably accurate" - what do you believe the margin of error on the population clock to be, and what is your basis for that belief? When the advertised accuracy of a figure is 0.01% and the discrepancy is 0.2%, that's highly significant; when the discrepancy is 500k and the source makes it clear that they don't know the margin of error, it's quite impossible to judge whether the discrepancy is meaningful.

"Those reports are used by every government and most non-government agencies, as well as international agencies. You won't find international agencies doing their own census" - However, as I explained above, Census figures are only the first step in quite a long chain of work to produce current population projections. In this case, as noted in the Factbook documentation, the CIA's figures begin with census data - undoubtedly sourced by the ABS - but after that, the ABS' US counterpart (Bureau of the Census) does its own projections - probably making a few different assumptions along the way. I'm sorry if the explanation is a bit dry, but if you want to understand why two figures might disagree, you need to know how those figures are calculated.

You've also ducked the issue of scope differences. The grey areas in defining who we count as the 'Australian population' amount to far more than 500,000 people, so even if by some miracle both ABS and BotC had perfect and current data for the entire world population, you can expect see discrepancies of this size depending on exactly who each organisation decided to count towards their totals. --GenericBob (talk) 01:39, 17 April 2009 (UTC)

Okay. So your argument for using 'European' as an ethnic category for Australia, in data derived from an ABS source, requires going to a different source, disregarding the categorisation used for 'Australia', and instead seeking out the data for a small offshore territory - one with a unique history that makes different categorisations relevant - and then taking this classification scheme back to the ABS data for Australia. With no assurances that the CIA calculations used the same methodology (and in fact some pretty clear evidence that they didn't, in that discrepancy for 'Asian' figures). And the CIA Factbook is untrustworthy in how it categorises Australia's ethnic groups because "what it says is wrong" (based on your comparison of a different data item to a primary source on another site), but the page on Norfolk Island is trustworthy - not just as a representation of NI, a small island of about 2000 people far out to sea, but also when you extrapolate its approach to a country of twenty million people. Do you really not see this as WP:SYNTH?

Reading broadly is great. Digging through sources to find the slightest hooks to support your case - while ignoring more prominent and relevant bits in those same sources that oppose it - is not. If you're serious about putting properly-sourced data in this infobox without OR... well, why not just copy the "ethnic groups for Australia" data across directly from the CIA Factbook? We're agreed that it's a good quality secondary source, and it specifically refers to 'ethnic' groups - we don't have the 'ancestry' issue to deal with. That would be far better than rummaging through the Factbook to find a reference to a small sub-entity of Australia and then transplanting the categorisation scheme to another source's data - unless the point of the exercise is to defend your previous edits rather than considering whether they're actually appropriate. --GenericBob (talk) 21:02, 16 April 2009 (UTC)

For those who aren't familiar with Norfolk Island, it's an external territory of Australia - while nominally part of the country, it's mostly self-governing (which is presumably why it has its own separate entry in the Factbook). A lot of its inhabitants are descendants of the Bounty mutineers, and there's a significant social distinction between them and much more recent arrivals from Australia and NZ - which is probably why the Factbook has chosen to divide it into "descendants of the Bounty mutineers, Australian, New Zealander, Polynesian". For choosing categories for an article on Australia, it seems more reasonable to go to the entry for something more representative for Australia overall... like, say, 'Australia'. --GenericBob (talk) 21:12, 16 April 2009 (UTC)

@AussieLegend, 1. the CIA world factbook doesn't list "Australian" as an ethnic group. 2. Encyclopedia Britannica may interchange the terms, but wikipedia does not. Removed Again Intranetusa (talk) 01:42, 20 April 2009 (UTC)

1. The CIA does indeed list "Australian" as an ethnic group. It doesn't matter that it's not listed next to Australia. It's listed somewhere in the Ethnic Groups column which means it's an identified ethnic group.
2. That's a very poor argument. Wikipedia relies on sources like Encyclopaedia Britannica for its information. If such a credible source interchanges the terms then there's no reason why we should ignore it unless we can find a more credible source that says Brittanica is wrong.
3. I notice that somebody reverted you. There's a message there. --AussieLegend (talk) 07:16, 20 April 2009 (UTC)

Seeing as that someone reverted with the comment "What does the CIA have to do with this?", I think the message is that it was a bad reversion by someone who was not aware of the talk page discussion. Intranetusa should have referenced the talk page, but still the revert was ignorant of this discussion. AKAF (talk) 07:30, 20 April 2009 (UTC)

You've selectively quoted from the edit summary, which actually states "You are being premature. This is being discussed on the talk page. And what on earth has the CIA got to do with it?". Quite obviously the editor was well aware of the discussion, despite your assertion to the contrary. --AussieLegend (talk) 07:50, 20 April 2009 (UTC)

The fact that the editor was unaware of the only valid source so far proposed for the ethnicity data makes me think that they were not up to date. AKAF (talk) 07:56, 20 April 2009 (UTC)

That isn't a fact - you're just making assumptions. You have no idea what the editor was or was not aware of. --AussieLegend (talk) 08:32, 20 April 2009 (UTC)

"Australian" is not an ethnicity, but neither is "White" or "Asian" or "Aboriginal" --they're all groupings of multiple races. The problem with categorising demographics in Australia is that a serious amount of people don't know where their ancestors came from or come from a mix of ethnic backgrounds... Can the words "ethnic groups" not simply be changed to something else (or does it come with the template thing)?

As for the sources, why is the argument about which is more credible even coming up? The ABS has actually conducted the research--it's a primary source and an internal agency, whereas the CIA (which is a foreign organisation based in the United States) is a secondary source that gets its info from other sources. Unless of course the postman lost my copy of the CIA's "tick the box next to your ethnicity" pamphlet... So if both sources are just as incongruous with the subject, why would you not automatically go with the primary source? That is if you can't change the wording of "ethnic groups"...

Night w (talk) 08:13, 20 April 2009 (UTC)

Wikipedia prefers secondary and tertiary sources. See WP:PRIMARY. I would point out that this is a policy rather than a guideline or essay. It is not up for discussion. AKAF (talk) 08:27, 20 April 2009 (UTC)

The key word there is "prefers". "Primary sources that have been reliably published (for example, by a university press or mainstream newspaper) may be used in Wikipedia, but only with care, because it is easy to misuse them" is what the policy actually says. I know you're going to bring it up so I'll preempt you. The policy also says "Any interpretation of primary source material requires a reliable secondary source for that interpretation." There was no interpretation. It was a simple summary of the data in the source based on geographical origin of the groups listed. Summarisation was required because there was too much information to place in the infobox, as I've stated previously and had to re-iterate far too many times. --AussieLegend (talk) 08:42, 20 April 2009 (UTC)

I think there is more than sufficient discussion above to establish that you have neither the training nor inclination to analyse that data source with sufficient care. It is further clear that you misused that data source (WP:SYNTH) by performing an analysis which was only weakly supported by the primary data. In contrast, the analysis of the CIA world factbook was performed by trained demographers, and is a good secondary source. This is a clear case where WP:PRIMARY applies, because the primary source does not say what you want it to, and WP:SYNTH does not allow an analysis of the sort you performed. I think it is extremely clear to everyone from the long threads above that the analysis you want to perform is far from simple. Further it is extremely questionable whether it is correct and whether you have the training to understand why it is not correct. I say this with the greatest of respect for your energy and other contributions, but the simple fact that you indicate the percentages to 0.01% accuracy is a very clear indication that you do not understand the data analysis at all. Therefore I would strongly suggest that in fact this is a textbook case of why WP:PRIMARY exists in the first place. AKAF (talk) 09:23, 20 April 2009 (UTC)

"have neither the training nor inclination to analyse that data source" - Exactly what part of "no analysis" and "simple summarisation" do you not understand?

"In contrast, the analysis of the CIA world factbook was performed by trained demographers" - It was? Who analysed it? As far as I can see you just duplicated what was in the entry. Is this your definition of analysis?

"WP:SYNTH does not allow an analysis of the sort you performed." - Again, for the umpteenth time, there was no analysis, it was a simple summary. There was a single source so, again for the umpteenth time, WP:SYNTH does not apply. - "Do not put together information from multiple sources to reach a conclusion that is not stated explicitly by any of the sources." - Note "multiple sources".

"I think it is extremely clear to everyone from the long threads above that the analysis you want to perform is far from simple." - You think that simple addition is not simple?

"the simple fact that you indicate the percentages to 0.01% accuracy is a very clear indication that you do not understand the data analysis at all." - Absolute and utter rubbish. This is yet another bad assumption. The only reason I used that precision is because I fully expected somebody would take issue if the percentages didn't add up to 100. I never once thought that somebody would take issue with people from Europe being labelled as European.

"Therefore I would strongly suggest that in fact this is a textbook case of why WP:PRIMARY exists in the first place." - I strongly suggest that this is a misinterpretation of policy. --AussieLegend (talk) 10:10, 20 April 2009 (UTC)

I am sorry that you appear to be having some trouble understanding the arguments of everybody else here. However I cannot be responsible for reiterating the same arguments for the nth time if you refuse to educate yourself on how data analysis, citation and demography work. I would also suggest error propagation and significant figures as further topics of study, and your local library should be able to help you with a good text. AKAF (talk) 10:57, 20 April 2009 (UTC)

"There was a single source so, again for the umpteenth time, WP:SYNTH does not apply." You have invoked both Britannica and the CIA Factbook to justify your analysis of ABS data. That makes three sources.

Have a look at the example of synthesis on WP:SYNTH: "(1)Smith claimed that Jones did X. Jones replied that X is acceptable... (2)The Harvard manual does not call X 'plagiarism'." The policy page makes it clear that this is synthesis - even though both statements are sourced, assuming that (2) is relevant to (1) is OR. "To make the second paragraph consistent with this policy, a reliable source would be needed that specifically comments on the Smith and Jones dispute and makes the same point about the Harvard manual and plagiarism. In other words, that precise analysis must have been published by a reliable source in relation to the topic before it can be published in Wikipedia by a contributor." This is synthesis even though the editor has merely implied the conclusion ("Jones is not a plagiarist") by sticking two sourced statements together.

Smith accused Jones of X. The Harvard manual does not call X 'plagiarism'. This, per example above, is OR.

Smith described this data as 'ancestry'. The Harvard manual defines 'ancestry' as equivalent to 'ethnic group'. Still OR.

The ABS described this data as 'ancestry'. Britannica defines 'ancestry' as equivalent to 'ethnic group'. Still OR.

The ABS described this data as 'ancestry'. Britannica doesn't actually define 'ancestry' as equivalent to 'ethnic group', but it uses it that way. OR as above, plus a layer of interpretation (and not a very safe one, at that - if you see a doctor smoking, would you take that as medical advice to go out and light up?)

The ABS described this data as 'ancestry'. Britannica doesn't actually define 'ancestry' as equivalent to 'ethnic group', but it uses it that way, so the ABS' data is equivalent to 'ethnic groups'. OR+ as above, plus a little bit more OR because now we're explicitly stating a conclusion from that synthesis instead of just insinuating it.

This is ABS data on ethnic groups. - OR++ as above. Leaving out the bits where the OR happened doesn't make it non-SYNTH, it just makes it badly sourced SYNTH.

"You think that simple addition is not simple?" - I'm not sure how you get percentages using only addition, but leaving that aside temporarily, I'm willing to acknowledge that the techniques you're using are simple in a couple of senses - they're naive, and they're easy to implement. The problem is that assessing whether they're correct is an entirely different question, and not a simple one at all. (But FWIW, the answer is an emphatic 'no').

If I tell you that Scot has three kids and Erin has two, and ask how many kids Scot and Erin have altogether, a primary schoolkid can calculate 2+3=5 - it's a very simple calculation. But that doesn't mean it's right... --GenericBob (talk) 03:05, 21 April 2009 (UTC)

"You have invoked both Britannica and the CIA Factbook to justify your analysis of ABS data. That makes three sources." - Wrong! Back in December, when I origianally summarised the information, I used a single source. The data has been modified since then, as different documents appeared at the ABS link but there was always only one source used for each change. The CIA Factbook wasn't referred to until 5 days ago in response to Internetusa's claim that Australian wasn't an ethnicity. Neither that or Britannica was ever referred to in the past and they were only used on 16 April to show that other reliable sources referred to those ethnicities, demonstrating that the usage is widespread. Good try but no cigar.

"I'm not sure how you get percentages using only addition" - There's some irony in the fact that before I settled on "The only reason I used that precision is because I fully expected somebody would take issue if the percentages didn't add up to 100", I had written, "The only reason I used that precision is because I fully expected some pedant would take issue if the percentages didn't add up to 100", but I thought pedant was too stong a word. I see now that changing the word was a mistake. Next time I'll stick with my gut. Percentage buttons are standard on even the most basic calculators today. Calculating percentages is no harder than simple addition. I can show you if you want.

"If I tell you that Scot has three kids" (etc) - Nice example but it's considerably more complex than summarising the data in the source and somewhat disingenuous of you to try using it as an example of something completely different. Again you've missed out on that cigar. --AussieLegend (talk) 05:52, 21 April 2009 (UTC)

Now please think about this, because I've put a lot of effort into a finding a simple example which explains the problems with your analysis of the data: If I tell you that at the zoo yesterday there were 100 monkeys, who consumed 50 bananas, 60 oranges and 20 watermelons for a total of 130 pieces of fruit. Assume that each monkey ate no more than one of each type of fruit. Your analysis method tells me that 38.46% of monkeys ate bananas, 46.15% of monkeys ate oranges and 15.38% of monkeys ate watermelons. There are several problems with this analysis:

1. Pieces of fruit is only weakly correlated with number of monkeys (as responses are only weakly correlated with persons)
2. You assume that the sets are independent (One monkey may eat more than one piece of fruit/one person may give more than one response)
3. At best it will be an approximation to reality, with an additional systematic error correlated with the size difference between the two sets (The size difference is 23% in this case, 17.6% for the census data)
4. It is particularly bad when you start to group sets. For instance to state that 53.85% of monkeys ate no oranges is a simple addition of the type which you have performed, but this has absolutely no basis in reality. Since each monkey ate no more than one of each type of fruit, then there are only 40 monkeys (40%) who actually ate no oranges. This is an error of 25.7% (of 53.85%).
5. The stated accuracy of the computation (plusminus 0.01%) is just plain wrong. For this computation a realistic number would be plusminus one monkey (1%).

Now please think about this, because I've put a lot of effort into a simple example which explains why your analysis method is flawed. AKAF (talk) 06:35, 21 April 2009 (UTC)

I had a longer response, but the editor ate it, which is perhaps just as well... while we could continue this argument until the end of time, I don't want to do that just for the sake of arguing. Whatever the problems may have been with content, I believe you (AussieLegend) have provided a solution in the form of a reliable secondary source that lists "ethnic groups of Australia" in a format suitable for the infobox, with no need for translation. You've also found an error in the ABS documentation, making it possible to get that fixed. And you've honoured your agreement that the methodology would have to change if the documentation changed. While I don't think we're going to reach agreement on the details of this particular issue, I would prefer to focus on a productive working relationship rather than continuing to thrash out an argument that's no longer contributing to the article. If we're all willing to accept the CIA Factbook figures here, I would be very happy to let the rest of this drop. Any objections? --GenericBob (talk) 07:47, 21 April 2009 (UTC)

I'm not entirely happy with the CIA data because I think it's problematic. In fact I'd rather see nothing there now but I'm not going to push for it. Let's leave things as they are and see what happens. --AussieLegend (talk) 08:16, 21 April 2009 (UTC)

If I could just return to the response to my talking about primary and secondary sources, if it's Wiki policy to go by secondary sources, why do the "ethnic groups" listed on the pages for Canada, New Zealand, the United States and the United Kingdom all go by statistics provided by their own government census bureaus? Is Australia so incompetent at gathering our own statistics that we have to rely on the "intelligence" of the CIA? Is that what you're saying? I don't like the CIA figures because I find them vague and offensive. I'm with AL on either changing them or removing them from the box entirely. Night w (talk) 08:20, 22 April 2009 (UTC)

It's not a question of competence. The ABS does not provide the statistics for the ethnicities in the form they are presented in the box (or noone has been able to find them). The data provided by the ABS requires further analysis. It is not a question of incompetence, but of preparation of the raw data. The analysis of demographic data is hard to do and requires trained people to do it, and the only source we could find which has done this analysis in the form which is useful for the infobox is the CIA world factbook. Aussielegend performed an analysis on the data which was provided by the ABS to get those older figures. This analysis has been shown by many people to be fatally flawed in a number of key areas.

Now, your dislike of the CIA aside, it does have competent analysts. I'm sorry that you prefer the outcome of Aussielegend's analysis, but his results were bollocks. I cannot speak for the other articles, but probably the relevant bureaus did the statistical analysis in the form required. If not then we prefer to use a good analysis, and saying that they used a poor analysis is no excuse for doing do here. Relevant pages are WP:OTHERSTUFF as referenced in WP:WAX.

If this really bothers you, why don't you write to the ABS and ask them to provide a good set of data for the Wikipedia infobox? This would actually be an optimal outcome, and maybe they can be persuaded. The data in that infobox is also very questionable in other areas, for instance the population is probably also not correctly cited (it is uncorrected population data), and the GDP comes from the IMF. AKAF (talk) 08:59, 22 April 2009 (UTC)

I haven't checked the others, but Stats NZ publishes ethnicity data (see e.g. [25]), based on a Census question that specifically asks "which ethnic group do you belong to?" (Note that their published percentages add to more than 100%, because they allow more than one response.) The Australian Census doesn't have an equivalent question; while I can think of at least six that are relevant to ethnicity (Indigenous identification, COB of person, COB of father, COB of mother, ancestry, and language), you'd have to put them together and apply some expert judgement in order to form them into ethnic categories. And as AKAF notes, you'd probably want to correct for nonresponse, since some groups have higher rates of nonresponse than others.

The fact that SNZ collects that information and ABS doesn't isn't about competence, it's about finite resources and priorities. The more data you collect, the more it costs to process; this caused major embarrassment in the 1976 Census, when the ABS got too ambitious and half the forms had to be shredded unprocessed. Even if you have an unlimited budget for processing, asking too many questions gets you to a point of diminishing returns (people get bored and don't finish the form). So everybody has to make hard choices about what to leave out. The NZ design is better suited to describing the country's ethnic groups, but it's weaker on other scores - for instance, it's easy to identify second-generation Australians, but AFAIK there's no way to pick second-generation New Zealanders. --GenericBob (talk) 12:40, 22 April 2009 (UTC)

Developed country?

The statement: "Technologically advanced and industrialised, Australia is a prosperous multicultural country and has good results in many international comparisons of national performance such as health care, life expectancy, quality-of-life, human development, public education, economic freedom, and the protection of civil liberties and political rights." is quite misleading.

First - Australia has very poor internet network comparing to many countries in the world. Secondly, a country with slow and in some instances still not electrified trians cannot be technologically advanced. Industraliazed also is misleading - a country that is predominantly a primary producer, that doesn't manufacture even one bicycle, TV etc - cannot be called industrialized.

Multicultural? - just because there are many migrants - it doesn;t mean it is multicultural - when de facto it is an English colony which enforces English language at schools and English way of life.

Freedom with internet censorship cannot coexist. Democracy while still monarchy? Do you remember Whitlam Government??? Is this a democracy when the monarch can dissolve a government within one day????

Health care with long queues to emergecy units, and long queues for basic public health services such as dentist (usually people wait 2-3 years) is not at the top of the world.

Life expectancy - especially is local Aboriginal communities is very low.

Human development - still no bill of rights.

Public education - very low, no foreign languages at school, ranking very low in maths among other countries.

Hope you can amend this sentence and be true about Australia —Preceding unsigned comment added by 115.129.17.160 (talk) 01:45, 19 April 2009 (UTC)

Woah--where do you live??? Australia has a higher percentage of Internet users per capita than Japan, the UK, the US, France... AWA Limited manufactures televisions and, I don't know about bicycles, but haven you ever heard of Holden?

The government doesn't enforce English culture--freedom of religion and society is guaranteed in the constitution and you only have to walk down a city street to notice multiculturalism. I live down the road from a mosque. And just because the GovGen can exercise dismissal powers (that have only ever been used in an emergency), doesn't mean she can impose direct rule--the people choose the government. On the Democracy Index, Australia is ranked among the top 10 most democratic countries in the world (alongside seven other monarchies) by The Economist.

I don't know where you line up for your health care provider, but is it not free? It's very expensive in most countries. Check this research, which ranks Australia's health care system above that of both Canada and the US. As for average life expectancy, Australia is ranked 5th by the UN. So it's not low "especially" in Aboriginal communities, just specifically in those communities.

Did you go to the "Developed country" page before you Wikilinked it for this section's title? Australia is ranked 4th...

We don't live in the Dark Ages down here...but I think you knew that, and were just ranting. If you're not happy with the level of freedom and development in Australia, then move to Scandinavia. Oh...but they're all monarchies....

Night w (talk) 07:29, 20 April 2009 (UTC)

Anon commenter, it might help to consider issues of geography. Australia has about the same population as Sweden, Denmark, and Norway combined, spread across roughly ten times the land area. This has a lot of influence on what's practical for infrastructure projects (e.g. rail and communications). --GenericBob (talk) 23:49, 21 April 2009 (UTC)

[removed remainder of thread, which had ceased to be even tenuously related to how to improve the article. Hesperian 04:42, 24 April 2009 (UTC)]

ref improve or remove

The following three lines are from the article:

• Australia willingly participated in World War I. [Ref: Bean, C. Ed. (1941). Volume I - The Story of Anzac: the first phase, First World War Official Histories, Eleventh Edition.]

Not supported by linked page, is Bean 1941 the reference? Conscription suggests that the inhabitants of 'Australia' were less enthusiastic!

• Many Australians regard the defeat of the Australian and New Zealand Army Corps (ANZACs) at Gallipoli as the birth of the nation—its first major military action. ["Rudd angered by Gallipoli remarks". BBC News. 1 November 2008. Retrieved 2 November 2008.]

This swiped text is from a BBC article that contrasts the views of a former Prime Minister with the current one. Should we add ",... and many don't"?

• The Kokoda Track Campaign is regarded by many as an analogous nation-defining event during World War II.

Uncited! Analogous?

– cygnis insignis 05:33, 19 April 2009 (UTC)

I added the appropriate cites from the recent scholarship, especially for Kakoda see Hank Nelson, 'Gallipoli, Kokoda and the Making of National Identity', J of Australian Studies online Rjensen (talk) 14:03, 19 April 2009 (UTC)

More likely, WP:FAR YellowMonkey (click here to vote for world cycling's #1 model!) 05:35, 19 April 2009 (UTC)

I agree with the 3 dot points. I also suspect that FAR may be in order. The risk is that once the FA status is questioned or removed, attempts to improve it could go the other wasy - ie, all bets are off and nothing is sacred. As major changes to the culture section over the last few months suggested, I suspect the article could lapse into a state of even less quality. At least the current FA status is a way to hold off decrease in article quality. Hmmmm --Merbabu (talk) 06:35, 19 April 2009 (UTC)

I disagree will all three dot points.

• Supporting participation in the war and supporting conscription are two entirely separate issues. It is possible, and indeed highly likely than many people who opposed conscription (including servicing soldiers who voted) still supported Australia's participation in the war, with volunteer enlistment. This is neither contradictory or unusual. The war effort was popular and bipartisan and tested through elections. Casting Australia as an unwilling participant is historical revisionism. -- Mattinbgn\^talk 01:27, 20 April 2009 (UTC)
• No, the fact that "many don't" is obvious from the "Many Australians..." rather than "Most Australians..." or All Australians ..." This is nit-picking.
• The Kokoda campaign was put forward by Keating as an new and better "Gallipoli" Sources for this should be easy to find. -- Mattinbgn\^talk 01:27, 20 April 2009 (UTC)

Let's not go too FAR :) There are many fine editors active there, but this is a the top priority article for a wikiproject with some of the best. cygnis insignis 13:09, 19 April 2009 (UTC)

Yeah, if preventative action is taken it will be all fine and dandy, like when I pre-emptively referenced flag of Australia last year. There is a problem that if people make a lot of ad hoc edits the prose can become disjointed likke a list of trivia though. YellowMonkey (click here to vote for world cycling's #1 model!) 01:13, 20 April 2009 (UTC)

anon edits

it seems every single one needs to be reverted because of vandalism. can we get this page locked again? Rehumanist - talk 12:58, 26 April 2009 (UTC)

29 percent of people in Australia Australian?

"29.06% Australian, 0.45% Australian Aboriginal, 52.8% European, 6.67% Asian, 4.69% Other, 6.33% Unstated[1]"?

Huh? How is 29 percent of the Australian population Australian?— Preceding unsigned comment added by [[User:{{{1}}}|{{{1}}}]] ([[User talk:{{{1}}}#top|talk]] • [[Special:Contributions/{{{1}}}|contribs]])

It doesn't say 29 percent of the Australian population is Australian. This section refers specifically to ethnicity. At the 2006 Census, 29% of respondents indicated that their parental ancestry was Australian, as indicated in the citation. --AussieLegend (talk) 06:54, 21 February 2009 (UTC)

I'm not sure that everyone taking the census properly understood the question at hand, then. Unless immigration of Australia is at a worldwide high, I find it hard to believe that half the population is directly European. This is where demographics becomes hard: were you born here, were your parents born here, were your grandparents born here? Sure, the caucasian population probably came from Europe at some time in the past 200 years, but when can a people start to identify themselves as being from where they are, i.e. when can these people identify themselves as 'Australian'. It seems hard to measure in this context, but maybe it's easier than I think. Maybe I misunderstood the question!Farglesword (talk) 21:56, 20 March 2009 (UTC)

The problem seems to be that some editors assume there is some scientific or rational basis for ethnic classifications. In fact ethnicity always has been based on self-identification, or arbitrary classification by others. The question asks for self-identity, which ia as valid a way to define ethnic origins as any other. --Michael Johnson (talk) 00:03, 21 March 2009 (UTC)

We have alota immigrants so australian would stand for the original settlers--Gcpeoples (talk) 05:15, 22 March 2009 (UTC)

I'd find it hard to believe that people didn't understand the question, which simply asked where their parents were born. 52.8% of responses said that one or both of their parents were born in a European country. That doesn't mean that the respondent was born in Europe so I don't see an issue with the figure. --AussieLegend (talk) 05:50, 21 March 2009 (UTC)

I think you've misunderstood the documentation here; there is no question on the 2001 or 2006 Census that asks for that level of data about where respondents' parents were born. (There's a copy of the 2006 form near the end of the online Census Dictionary, and the 2001 form is here.)

Questions 14 and 15 (13 and 14 on the 2001 form) ask "Was the person's father/mother born in Australia or overseas?" There is no option to indicate where overseas; in this data, 'country of birth of parents' is provided only at the level of 'Australia' vs 'rest of world'.

Question 18 asks "What is the person’s ancestry?" The question itself doesn't mention parents, but it points respondents at page 7 of the Census Guide for more information. That section states: "...Ancestry is not necessarily related to the place a person was born but is more the cultural group that they most closely identify with. For example, a person may be born in New Zealand but have Samoan ancestry." While it advises respondents to "consider the origins of the person's parents and grandparents for example", it certainly doesn't lock 'ancestry' in with country of parents' birth - indeed, it encourages people who are descendants of indentured South Sea Islanders to describe themselves as 'Australian South Sea Islander', even though those ancestors would have arrived in Australia about 100 years ago.

The source we're looking at, "Ancestry by Country of Birth of Parents " combines data from these three questions. However, since the analysis used to populate Australia's infobox only uses the 'total' column, it ignores all information supplied on questions 13 & 14; the only thing it represents is responses to question 18. Which doesn't "simply ask where their parents were born", not by a long shot. (Apologies for the late response to this point, but this talk page is getting pretty bloated and it's easy to miss stuff.) --GenericBob (talk) 00:48, 16 April 2009 (UTC)

And where did you get that number from, AussieLegend? I've looked at the data, and only 9, 000 Australians identified as 'European'. Most Australians of European ethnic origin identify as German, English, Italian, etc. not as just 'European'. Besides, Anglo-Celtic people that identified their ancestry as English, Scots, Irish, etc. do not form a distinct ethnic group from those who identified their ancestry as just 'Australian', and the field clearly says 'ethnic groups' not 'self-identified ancestral identity'. Aren't many European Australians of more than one European ancestry, anyway? Does the Census allow more than one answer to the "ancestry" question? Saimdusan Talk|Contribs 05:15, 1 April 2009 (UTC)

Also, where did the number for 'Asian' come from? Does it include subcontinental Indians? Does it include Middle Easterners? If it includes Middle Easterners, does it include Egyptian Arabs? If it does not, why would Lebanese Arabs, who are more related to Egyptian Arabs than to Chinese, be included in this category? Saimdusan Talk|Contribs 05:17, 1 April 2009 (UTC)

I am going to remove the ethnic groups field from the article for the time being, as how it stands now it just serves to confuse and would better be served in a section rather than just "29% Australian, 53% European, 7% Asian, etc.". It should however be added again once this issue is solved. Saimdusan Talk|Contribs 05:19, 1 April 2009 (UTC)

The data came straight from the Australian Bureau of Statistics, as indicated in the citation provided in support of the data. The ethnicities listed are summaries based on the origin of the parents, ie parents from England, Gemany, Italy, Greece etc are listed as European, Chinese, Filipino, Vietnamese are listed as Asian etc., as per convention. I really don't know where you got the rather ridiculous figure of 9,000 Europeans. There were 13,427,947 responses indicating that one or both parents were from European countries. That's a tad higher than 9,000.

There should be no confusion at all, especially if you take the time to read the citation provided. Since the information is cited and since the citation is from a reliable source I've restored it. If you can provide a better source than the ABS for the data then there might be reason to remove it, but until then it should stay. Don't delete cited information just because you didn't bother to check the citation. --AussieLegend (talk) 05:51, 1 April 2009 (UTC)

So, I did take the time to read the cited document. In the first part (2001 census) it states:

• Both parents born overseas: Responses (6,027,352) Persons (5,410,162)
• Father only born overseas: Responses (1,849,204) Persons (1,245,494)
• Mother only born overseas: Responses (1,214,965) Persons (815,324)
• Both parents born in Australia: Responses (12,258,949) Persons (9,983,897)
• Country of birth not stated: Responses (1,218,001) Persons (1,133,431)
• Total responses: Responses (22,568,471) Persons (18,588,308)

Giving the following percentages for 2001:

• Both parents born overseas: 26.71%/29.11%
• Father only born overseas: 8.19%/6.7%
• Mother only born overseas: 5.38%/4.39%
• Both parents born in Australia: 54.32%/53.71%
• Country of birth not stated: 5.4%/6.1%
• Total responses: 100%/100%

For 2006 it looks pretty similar

• Both parents born overseas: 26.51%/29.56%
• Father only born overseas: 8.32%/6.55%
• Mother only born overseas: 5.66%/4.43%
• Both parents born in Australia: 53.07%/51.79%
• Country of birth not stated: 6.44%/7.68%
• Total responses: 100%/100%

I'm not sure why the responses and persons are not the same, so I've given both numbers. It looks like the table in the article has either switched the numbers from the European and Australian parts, since its hardly possible for both parents to have been born in Australia and be European or its using the 13 Million responses you cite. This is the sum of the self-described ethnicity of the respondent, rather than where the parents were born. The citation is quite confusing though. Particularly since "Australian" isn't really a recognised ethnicity. Probably it would be better to group "Australian" with "other". AKAF (talk) 11:21, 1 April 2009 (UTC)

Firstly, there may be discrepancies between the figures in the article and what I've written above, and what I'm about to present because, once again, the ABS has changed the document linked to at the url in the citation. This was previously discussed and the article updated to reflect the more detailed data in the newer document.[26] The url is now linking to the older document again so I'll refer to the figures in that document in the following paragraphs, rather than the document that was used for the article.

The figures in each column are responses, not people. That can be confirmed simply by adding the columns and comparing the results to the two totals for each column. The only place that persons are used is in the totals section and the persons totals don't relate directly to the totals for individual ethnicities. In the right-most column (Total), there were 6,674,946 responses who identified as Australian out of a total of 22,568,471. That represents 29.57% of the total responses. The European ethnicities (Croatian, Dutch, English, French, German, Greek, Hungarian, Irish, Italian, Macedonian, Maltese, Polish, Russian, Scottish, Serbian, Spanish and Welsh) represent a total of 11,798,539, or 52.278%.

The purpose of the infobox information is to show the breakdown of the country and I think there would be more confusion if "Australian" was not identified as an ethnicity or if Australian was lumped in with "other". Either of these would give the impression that there were no people identifying themselves as Australians living in Australia and that's just plain silly. --AussieLegend (talk) 12:20, 1 April 2009 (UTC)

Yeah, the document isn't easy to read. I think the thing which people are objecting to is to referring to "Australian" as an ethnicity rather than a nationality. Certainly over half of all respondents would refer to themselves as Australian, since both parents were born in Australia. Its a bit pointless at some juncture to cite the document, when fully 1/3 of all respondents effectively have no ethnicity (~36%=Australian plus no comment). That gives an error margin of +60% on the figure for European ethnicity and +600% on the figure for Asian ethnicity. With these kinds of margins for error it would be better to leave out the data or look for a better source. AKAF (talk) 13:59, 1 April 2009 (UTC)

I think the problem here is that ethnicity is something that is difficult to define, and that self-definition is about the only reasonable way to go about it. There is no "scientific" definition of ethnic origin, and to try and impose one is to start on a slippery slope towards racism. So if people want to identify themselves in an ethnic way as "Australian" good luck to them. And no I don't think we should assume anything about what that means (ie we can't assume such people are of European decent, for instance). --Michael Johnson (talk) 23:44, 1 April 2009 (UTC)

I agree. Self identification explains why there are some peculiar figures in the source document, such as 1,455 of the 93,486 respondents who identified themselves as "Australian Aboriginal" stating that both parents were born overseas. --AussieLegend (talk) 06:00, 2 April 2009 (UTC)

The point which I was trying to make is that, although the reference exists, the quality of the data is dubious at best. I don't see that its so important that the article would fail without this piece of information. It's not like the article is struggling for length, so I would suggest removing the data from the infobox. The data needs more context, which can be better provided in the text. AKAF (talk) 06:48, 2 April 2009 (UTC)

"quality of the data is dubious at best". Do you realise that the Australian Bureau of Statistics is a Australian Government body (Similar to the United States Census Bureau) meaning that it's the most reliable source that you're going to get and find. Just because you think it's wrong that it should be removed. If you don't agree with the figures then take it up with the ABS. Bidgee (talk) 08:13, 2 April 2009 (UTC)

The data is from a very reliable source so it's not really up to us to dispute it unless we have citations from more reliable sources to the contrary. Infoboxes are supposed to summarise information from the main document. There should be discussion on it in the prose that gives it context anyway. Deleting it from the infobox just because somebody doesn't understand it is not appropriate. If we were to delete information for that reason, most infoboxes would be empty. --AussieLegend (talk) 07:02, 2 April 2009 (UTC)

A reliable source can, and does, produce unreliable data when the signal to noise ratio is high. I'm particularly commenting that its pretty brainless to quote this source as reliable for the figure of 6.67% Asian when 6.33% left their ethnicity unstated and 26% refused to comment, and left their ethnicity as "Australian". In this case the "true" figure for asian ethnicity could well be 12% or 20% and be fully consistent with the data. If we were honest in quoting the data accurately we would use error bars and divide the nonspecific data between the specific respondents, using the assumption that no specific group refused to answer this question (which assumption is invalid, but anyway). In this case the data would look something like this:

1. 0.7% $\pm$ 0.2% Australian Aboriginal
2. 81.7% $\pm$ 28.9% European
3. 10.3% $\pm$ 3.7% Asian
4. 7.3% $\pm$ 2.6% Other

The values should definitely only be quoted to one decimal place, and probably better to one significant figure. As they stand, the values are a red flashing sign that the writers of the article don't understand data analysis. Please understand that these figures are no more original research than the numbers currently in the article, since "percent ethnicity" is not a component of that document. AKAF (talk) 13:11, 2 April 2009 (UTC)

There's really not much that can be done if people don't wish to state their ethnicity. All you can do is to include the unstated figure in the list, so that the reader realises that there is a margin of error in the figures, and that has been done. I'm not sure what you mean about 26% percent refusing to comment. Only 5.6% (using the figures in the current source) refused to identify an ethnicity. As for identifying as Australian, it's already been said that there's nothing wrong with that. My parents and grand-parents were born in Australia and I identify as Australian. As far as I'm concerned that's my ethnicity. I'm sure many feel the same way. Your claims regarding percentage errors do constitue original research since they're not explicitly stated in the document and you need to make assumptions to arrive at the errors that you've presented. The figure are stated, which is why we use them. Working out percentages (not percentage errors) based on the stated figures is not original research, because no assumption is necessary. Anyone reading the document can work out that 6,674,946 is 29.58% of 22,568,471 without making a single assumption. --AussieLegend (talk) 13:54, 2 April 2009 (UTC)

The errors, sure. I mean the percentages calculated in the way above. Particularly since the percentages in the article are not quoted from the document, but are assembled from the document. AKAF (talk) 14:29, 2 April 2009 (UTC)

Of course they refused to state an ethnicity, AussieLegend. Once again, 'Australian' is not an ethnicity. White Australians who identified as Australian are not distinct from White Australians who identified as some European. I got the 9,000 figure from the data itself, the data says that 9,000 people identifed their ancestry as 'European'. Of course, it also says that there are millions of people identifying with particular European peoples. But pulling this new meaning out of the census data is WP:OR. The Aborigines are the true 'ethnic Australians' and Anglo-Celtic Australians should not perpetuate this false hierarchy with whites/aussies at the top, wogs/ethnics in the middle and abos/blacks at the bottom. I can understand seeing a second-generation paki (Punjabi) and wog (Serb, Slovene, Russian) such as myself as non-Australian but definitely Indigenous peopple are Australians! The best thing to do would be to simply show the ancestry data as it is - show the highest reported ancestries (rather than this WP:OR 'European' identity that I as a European Australian actually do support), then add a note for 'Australian'. Fixed, simple enough. Saimdusan Talk|Contribs 04:59, 3 April 2009 (UTC)

Once again, 'Australian' is not an ethnicity Why not? It seems to me to be an act of arrogance, or indeed racism, to try and appropriate people to ethnic groups they don't identify with. There is no scientific basis to ethnic definitions, only historical, cultural, political and geographical definitions, and these are both variable and numerous. So there are no hard and fast rules regarding ethnicity. If people identify as "Australian" that is what they are. And once again, no assumption can be made that they are necessarily of European decent. --Michael Johnson (talk) 23:21, 3 April 2009 (UTC)

Identifying as Australian is not refusing to state an ethnicity. It's stating that your ethnic group is Australian. The makeup of ethnic groups evolves over time. Ethnic groups aren't created one day and then stay the same for ever. People in England, many of whom descended from Roman occupiers of the British Isles aren't classified as Italian. They're classed as English. It's a similar story in Australia (and everywhere else). My family's roots in Australia date back to the First Fleet so I regard myself as Australian, not European, despite the fact that my ancestors are originally from there. Your definitions of various ethnic groups certainly seem to have some racist undertones.

As for your mythical figure of 9,000, looking at the citation I see no evidence supporting that. The smallest figure listed is 43,831 and that's for Americans.

That leaves the rather silly claim that identifying people from Europe as European is WP:OR. I'm really not sure what to say about that one. --AussieLegend (talk) 07:13, 3 April 2009 (UTC)

You had 512 relatives on the first fleet? Congratulations! Would it not be better to use the data for where the parents were born instead of the self-reported ethnicities? This is probably significantly more reliable and gives a more solid (but different) demographic picture. Just put the percentages for whom one or more parent was born in $COUNTRY. Improves the quality and still gives an idea of the immigrant makeup. For a table you don't want any data which requires multiple minutes of analysis to understand. It's far better to only report facts for which there is high-quality data. Ethnicity data will always be dubious, and no amount of massaging will help that. AKAF (talk) 07:44, 3 April 2009 (UTC)

"You had 512 relatives on the first fleet?" - What the hell are you talking about?

"Would it not be better to use the data for where the parents were born instead of the self-reported ethnicities?" - Because that isn't identified in the source data. "Overseas" is not an ethnicity. --AussieLegend (talk) 07:58, 3 April 2009 (UTC)

I was simply pointing out that you have >512 ancestors who were of reproductive age in the 1790s, to say that your ethnicity is wholly defined by one of them is the height of craziness. Many Australians have at least one Aboriginal relative (mostly this was hidden by the family) but this is not a defining trait of ethnicity, if it was more than 100 years back. The table indicates where the parents of the respondents were born. Is there another way to read it? AKAF (talk) 11:45, 3 April 2009 (UTC)

Once again, ethnicity is self defining. If people identify as Australian, it really doesn't matter if there ancestors arrived 30,000 years ago, 200 years ago, or yesterday. That's what they are. You're welcome to produce a source that argues differently. --Michael Johnson (talk) 23:21, 3 April 2009 (UTC)

Arbitrary section break

1. Even Anglo-Australian assimilationist ethnic nationalists do not see themselves as being of Australian ethnicity. They (and presumably you, AussieLegend, do too) see themselves as "white".

2. In this context, claiming that people from Europe are European IS original research. The source that is cited does not mention the word "European". Not once. If one looks at the full classification list by sex, 9, 000 people identified their ancestry as European. AussieLegend, what is your definition of European? More importantly, why should we go by your definition? Does 'European' include Turks? Why/why not? Aren't some Turks 'from' Asia? But how can one know which Turks are 'European' and which 'Asian' based on the data? Why 'European'? Why not Eurasian, or for that matter, Afro-Eurasian? Why not western/northern/eastern/southern European? 3. The cited data shows responses to the question "What is your ancestry?". The data does not mention "ethnicity". 4. Where did you get the data for 'European' anyway? Did you lump all of the 'European' respondents together? You do realize that one can select two ancestries on the census, right? This would mean that the data overlaps, so they can't be tallied together. 5. Why did you put 'Australian' respondents first? Europeans were larger. Saimdusan Talk|Contribs 06:11, 5 April 2009 (UTC)

1. "Anglo-Australian assimilationist ethnic nationalists" - Did you get that from a buzzphrase generator? I, and I'm sure most people, have no idea what that means. As for being white, what does that have to do anything?

2. I think you completely misunderstand the concept of original research, a policy that states, "Wikipedia does not publish original research or original thought. This includes unpublished facts, arguments, speculation, and ideas; and any unpublished analysis or synthesis of published material that serves to advance a position." Claiming that somebody from Europe is European is none of those things. Quite simply, since you seem to be having trouble with this, somebody from the continent of Europe is European. It's a well known, uncontroversial fact. It's somewhat ironic that you've asked what definitions I've used since there are wikilinks to the appropriate articles right next to the figures presented. I've just confirmed them for your benefit. I really don't see why you should have an issue with this. The information in the source document has been summarised into European, Asian etc because, as I've already stated, the infobox is a summary of what's in the article. Detailed lists shouldn't be included. When we did include more detail, there was an issue with people who couldn't understand that there are more Germans in Australia than Greeks now.

3. I suggest that you click on the "ANCESTRY" link in the source. If you do, you'll see that the ABS believes that "a person’s ancestry, when used in conjunction with the person’s country of birth and whether the person’s parents were born in Australia or overseas, provides a good indication of the ethnic background of first and second generation Australians". (bolding added for emphasis) Granted, the source includes everyone, but that's really not an issue.

4. You are clearly missing the point. The first and last columns are all that are relevant. The first column identifies the respondent's ethnicity, the last identifies the number of respondents who identify with that ethnicity. Where the respondent's parents were born is not important. What is important is the ethnic group with which the respondents identify.

5. Because this is an article about Australia, so Australian seems the most relevant ethnicity. The source document does the same. If this was an article about India, I'd put Indian first. It's not rocket science. --AussieLegend (talk) 07:26, 5 April 2009 (UTC)

" Anglo- Australian assimilationist ethnic nationalists " -- or ' Bogan ' Lejon (talk) 05:04, 6 April 2009 (UTC)

Yes, ancestry is an indicator of ethnicity, but that doesn't make them the same thing. The source document describes the data as 'ancestry', and in other edits you've been punctilious about sticking to the exact wording used by the ABS; let's apply the same standard here. You will notice that the Census question which provided this data refers only to 'ancestry' and doesn't refer to 'ethnicity' at any point. --GenericBob (talk) 05:30, 8 April 2009 (UTC)

Further, the percentages given in this article are not taken from the cited source; they're derived from it, apparently by dividing the number of responses in each category by 'total responses'. This is misleading, since it gives readers the impression that only 29.6% of respondents identified ancestry as 'Australian', when in fact 35.9% did so (6,674,946 of 18,588,308). --GenericBob (talk) 05:45, 8 April 2009 (UTC)

The figures are derived by simple mathematics, directly from the source data. Anyone, well almost anyone it seems, can verify the results. The figure of 6,674,946 that you've quoted is the figure for 2001, not 2006. The 2006 figure is is 7,371,824. In any case, your result is in error for 2001 because you've divided by the number of persons rather than by the number of responses. You've divided the number of apples on the bench by the number of oranges that you've eaten to determine the number of grapefruit left on the tree. That's why you've arrived at 35.9% for 2001, instead of the correct figure of 29.57%. Just to make it clear for anyone else who may look at the wrong data, the relevant cells are A56:A88 (ethnic group), G56:G88 (number of responses for the corresponding cells in column A), and G89 (Total responses). Eveything else is irrelevant. --AussieLegend (talk) 07:29, 8 April 2009 (UTC)

I was using the 2001 figures because those were the ones that you put in the article, and the percentages given were derived from those 2001 figures. Now that we've updated to 2006 figures, I'm happy to discuss those; the numbers are slightly different but the problems are unchanged.

The first is representing the data as 'ethnic groups', when the source (and the Census question on which that data is based) do not make any such representation. It's certainly related to ethnicity, but that doesn't make it the same thing. The ABS' notes on ancestry specify that it is a "good indication" of ethnic background of first and second generation Australians, when used in conjunction with the person's country of birth (which is a separate data item - Q12 on the Census form). Neither of those conditions apply here.

As an illustration of why terminology is important here, compare the responses to this question with the question on Indigenous identity. When asked "is the personal of Aboriginal or Torres Strait Islander origin?" (Q7), 425,511 people ticked 'yes' to Aboriginal ancestry (that includes 17,811 who also ticked TSI)[27]. But when asked about their ancestry (Q18 on the same form), only 115,820 people identified Aboriginal ancestry. At first glance, either of those two do look like they should be 'close enough' to ethnicity - but they disagree by half an order of magnitude! When the response is so sensitive to the way the question is worded, it's vital that Wikipedia represents that wording accurately. We should not be talking about 'ethnic group' when the source is based on 'ancestry'.

The second is that dividing by number of responses is inappropriate. When a reader sees "51.05% European, 6.33% unstated", the natural interpretation is that 51.05% of Australian people answered 'European', and 6.33% didn't answer - and this is incorrect. In fact, about 8.1% of respondents didn't answer at all. At the point where you decide that it's more appropriate to divide by responses than people, you're engaging in interpretation, and we get into the realm of original research.

I would appreciate it if you'd refrain from snideness; it's not conducive to amicable editing. And please don't remove accuracy tags until the issue has been resolved. --GenericBob (talk) 01:02, 9 April 2009 (UTC)

Dividing by the number of responses is entirely appropriate because all of the numerical data in rows 18-51 and 56-89 are numbers of responses, not numbers of people. There are only 12 cells (B52:G52 and B90:G90) that refer to persons. There is no explanation in the source of how the figures in rows 52 and 90 were determined, only an explanation of why they do not correspond the the numbers of responses (cells A97 and A98). If you tally the figures in each column you will see that the totals correspond to the values in rows 51 and 89, not 52 and 90. You CAN NOT divide the number of responses by the number of persons, as you did to arrive at the incorrect figure of 8.1% instead of the correct figure of 6.33% because, as is stated in the source, the sourcee "is a multi-response table, and therefore the total responses count will not necessarily equal the total persons count." Dividing the number of responses by the total number of persons is not just OR, it's WP:SYNTH. --AussieLegend (talk) 08:40, 9 April 2009 (UTC)

Can you explain why all other computations on the data are WP:SYNTH while your computation is not? AKAF (talk) 09:08, 9 April 2009 (UTC)

More precisely: Could you please provide the row and column number where in the source it is explicitly stated that 51.05% of Australians are European, as required by WP:SYNTH? AKAF (talk) 09:11, 9 April 2009 (UTC)

WP:Primary sources states: All interpretive claims, analyses, or synthetic claims about primary sources must be referenced to a secondary source, rather than original analysis of the primary-source material by Wikipedia editors. It then names census results as an example. In the preface to the ASC:CEG

It is difficult to define ethnicity in a way that is both useful and generally acceptable and it is not the function of this document to attempt an extensive definition of the concept. However, because the words ‘ethnicity’ and ‘ethnic’ are associated with many different meanings it is useful to provide some definitional material. The Macquarie Dictionary (Third Edition, 1997) provides the following meanings:

1. relating to or peculiar to a population, especially to a speech group, loosely also to a race. 2. relating to the origin, classification, characteristics, etc., of such groups. 3. of or relating to members of the community who are migrants or descendants of migrants and whose native language is not English. 4. recognisable as coming from an identifiable culture.
For the purposes of ASCCEG it is sufficient, and not controversial, to say that the term ‘ethnicity’ refers to the shared identity or similarity of a group of people on the basis of one or more factors, including the following which were enunciated in a report entitled The Measurement of Ethnicity in the Australian Census of Population and Housing [Borrie report, 1986]

... Since the publication of the Borrie Report in April 1984, the multicultural nature of Australian society has further developed resulting in a more sophisticated and enlightened approach to cultural diversity. Nevertheless, ... Dennis Trewin ASC:CEG, 2000-01

It is appropriate to challenge any information interpreted or extrapolated from data in a primary source, even if it was given some context in the article. The data was collated and massaged, apparently, "to satisfy wide community interest in the ethnic and cultural composition of the Australian population and the characteristics of particular migrant community groups.", and it is tempting synthesise that interest with Australia's historical (or hysterical) xenophobia and obsession with race. One need look no further than earlier Australian censuses to see 'arbitrary sectarian breakdowns', to put a twist on this section's heading. Inclusion of the breakdown in the article would be highly questionable, requiring secondary sources, it is unacceptable to thrust it upon a reader as infobox content. The data presented in these garish oversimplifications is, at best, next to useless as encyclopaedic information. The 'ethnic' section of this infobox, in a featured article, is effectively operating as a POV and OR fork and should be removed.

It should be obvious to anyone that dividng one type of figure (responses) by another (persons) to come up with a third type (percentage) is completely wrong when the source document says that the two don't correlate. In order to determine the percentages you need to be using the same data, which I am. As I already explained above, "There is no explanation in the source of how the figures in rows 52 and 90 were determined, only an explanation of why they do not correspond the the numbers of responses". The figures for persons have come from somewehere else, even though they are included in the source used. It is not possible to arrive at the figures presented for persons using the data contained in the source. This is not the case when determining percentages based on responses. Again, as I've already stated above, "The figures are derived by simple mathematics, directly from the source data. Anyone, well almost anyone it seems, can verify the results". WP:SYNTH does not apply because only one source is being used for the calculations, which you can confirm yourself, unless you have absolutely no idea where Europe and Asia are. If that's the case you can always follow the links that are provided. --AussieLegend (talk) 09:43, 9 April 2009 (UTC)

AussieLegend, it might seem sensible at first to divide responses by total responses. However, it is complete nonsense to put percentages derived this way in the infobox. These percentages are the percentages of responses in each category, but a normal reading of the infobox would suggest that it is giving percentages of people identifying with each ethnicity. It is not possible to do this, since while the ABS has made sure people are only counted once in each of their categories, there are sure to be quite a few counted twice in the supercategories listed in the infobox. Summarising data should not be covered by WP:SYNTH, but any suggestion that this sort of summary of these response figures give percentages of various ethnicities is simply wrong. JPD (talk) 11:38, 9 April 2009 (UTC)

If the ABS has counted each person once in each of the categories, and the sum of the categories in which each person is counted once is the total of each supercategory, how can each person be counted more than once in the supercategory but not in the individual categories? That's like saying 1+1+1+1+1=6. I know what you're getting at but the error that may exist is so low as to be inconsequential. The figures are really a guide anyway, since they are based on data that is now 2.5 years old. Census data is only accurate on the day of the census. I can't see why people are getting their knickers in such a twist. According to Australia's population clock, we have 1,871,101 more people in the country, a 9.4% increase, and we have no idea what ethnicity they are. There's bound to be a big dent in the figures that anyone here could calculate. As I said, it's just a guide. --AussieLegend (talk) 12:35, 9 April 2009 (UTC)

Yes, supercategories are a problem - there is no way to tell how many distinct people identified ancestry somewhere in Europe. The data you are using is not suited for the use to which you're trying to put it, and in attempting to pound a round peg into a square hole you're engaging in OR. --GenericBob (talk) 14:49, 9 April 2009 (UTC)

You're under a misapprehension. The fact that there isn't a one-to-one match between 'responses' and 'persons' does not mean we can't meaningfully divide them. Let's try a simpler example: I ask Bob and Jane to pick up to two words that describe their cars. Bob picks "red" and "Ford"; Jane picks "red" and "Holden". Two people, four responses (two "red", one "Ford", one "Holden"). So how would we represent this in the Bob and Jane's Cars Infobox?

Under your approach - dividing responses by total responses - we would represent this as "50% red, 25% Ford, 25% Holden". Under mine, it would be "100% red, 50% Holden, 50% Ford" - which is a rather more accurate representation of the situation. There is no problem with dividing a count of responses by the count of persons when what you're looking for is the proportion of persons who gave a particular response.

And yes, my approach most certainly is OR. I believe it's a higher quality of OR - I have a fair bit of professional familiarity with this sort of work - but it certainly doesn't belong in article space. Nor does yours. The mathematical side of the analysis is OR, because of this issue of interpretation, and conflating 'ancestry' with 'ethnic group' is also OR. --GenericBob (talk) 11:58, 9 April 2009 (UTC)

Under your approach the breakdowns total 127.98%. Put that in the infobox and see what reaction you get.--AussieLegend (talk) 12:35, 9 April 2009 (UTC)

This is only a problem if you believe these categories are mutually exclusive - in which case, you're in trouble, because the question used to collect them is specifically designed to allow for a person falling into more than one category. In any case, as I've already said above, I have no intention of putting them in the infobox. You seem to have decided that this infobox MUST be filled, and that one way or another the ABS 'ancestry' data is an appropriate source with which to achieve that; both of these are bad assumptions. --GenericBob (talk) 14:49, 9 April 2009 (UTC)

There's no analysis of the data involved. It's just a summary of the presented data in an easier to read format than:

• • Australian - 7,371,824
  • Australian Aboriginal - 115,280
  • American - 56,283
  • Chinese - 669,890
  • Croatian - 118,046
  • Dutch - 310,089
  • English - 6,283,650
  • Filipino - 160,374
  • French - 98,333
  • German - 811,540
  • Greek - 365,147
  • Hungarian - 67,628
  • Indian - 234,722
  • Irish - 1,803,741
  • Italian - 852,418
  • Lebanese - 181,745
  • Macedonian - 83,983
  • Maltese - 153,803
  • Maori - 92,912
  • New Zealander - 160,681
  • Polish - 163,802
  • Russian - 67,054
  • Scottish - 1,501,201
  • Serbian - 95,365
  • Sinhalese - 73,856
  • South African - 79,521
  • Spanish - 84,322
  • Turkish - 59,393
  • Vietnamese - 173,658
  • Welsh - 113,250
  • Other(c) - 1,397,647
  • Ancestry not stated - 1,609,443
  • Total Responses(d) - 25,410,601

which doesn't fit in an infobox too well. --AussieLegend (talk) 09:51, 9 April 2009 (UTC)

No, it doesn't. Neither would the explanation of how they arrived at those figures. 10:10, 9 April 2009 (UTC)

• In order to prove your claim that you have performed no synthesis from the data, could you please provide the row and column number where in the source it is explicitly stated that 51.05% of Australians are European. This is not visible in the excerpt which you quote above. AKAF (talk) 11:21, 9 April 2009 (UTC)

I agree that trying to fit all that into an infobox would be a horrendous mess. The correct solution is not to violate WP:OR, but to leave that field empty until such time as an appropriate source can be found. --GenericBob (talk) 11:58, 9 April 2009 (UTC)

After checking the ABS site in search of a more appropriate source I believe we actually have one already. WP:OR has not been violated. There should be no problem. --AussieLegend (talk) 12:41, 9 April 2009 (UTC)

Please provide the row and column number where in the source it is explicitly stated that 51.05% of Australians are European. This is not visible in the excerpt which you quote above. Doing this will demonstrate that your synthesis is not WP:OR. AKAF (talk) 10:06, 10 April 2009 (UTC)

How many times does this question have to be answered before you stop acting like a broken record? The data presented is simply a summary from a single, cited source, directly converted to a percentage for ease of comparison so WP:OR doesn't apply. --AussieLegend (talk) 11:45, 10 April 2009 (UTC)

No, it's not 'directly' converted to a percentage - before making that conversion you've converted the data into supercategories e.g. 'European'. Given the multi-response nature of the data, the assumption that it is valid to sum individual categories to get a meaningful supercategory is not a trivial one, which makes it OR. Choosing between dividing by number of responses or number of persons is an analytical decision, and that makes it OR. Interpreting 'ancestry' as equivalent to 'ethnic group' - also OR.

Since this is chewing up a lot of space on the Talk page and it doesn't look like we're getting any closer to agreement, I'm going to put this up for RFC. --GenericBob (talk) 13:47, 10 April 2009 (UTC)

Stating thst it's "not 'directly' converted" is being extremely pedantic. The numbers are just being added based on continental location for ease of display. --AussieLegend (talk) 13:57, 10 April 2009 (UTC)

Second arbitrary section break

Since you haven't answered any of my questions, I'll just ask them again.

AussieLegend, what is your definition of European? More importantly, why should we go by your definition? Does 'European' include Turks? Why/why not? Aren't some Turks 'from' Asia? But how can one know which Turks are 'European' and which 'Asian' based on the data? Why 'European'? Why not Eurasian, or for that matter, Afro-Eurasian? Why not western/northern/eastern/southern European? 3. The cited data shows responses to the question "What is your ancestry?". The data does not mention "ethnicity". 4. Where did you get the data for 'European' anyway? Did you lump all of the 'European' respondents together? You do realize that one can select two ancestries on the census, right? This would mean that the data overlaps, so they can't be tallied together.

1. "Anglo-Australian assimilationist ethnic nationalists" - Did you get that from a buzzphrase generator?

Anglo - Of or relating to England or Britain. Australian - Of or relating to Australia. Therefore, an Anglo-Australian is an Australian of British/English ancestry. Assimilationist - a proponent of assimilation. Ethnic nationalism - an ideology where the nation is defined by ethnicity.

As for being white, what does that have to do anything?

Oh come on, "whiteness" is one of the main factors of European ethnic identity.

I think you completely misunderstand the concept of original research, a policy that states, "Wikipedia does not publish original research or original thought. This includes unpublished facts, arguments, speculation, and ideas; and any unpublished analysis or synthesis of published material that serves to advance a position." Claiming that somebody from Europe is European is none of those things. Quite simply, since you seem to be having trouble with this, somebody from the continent of Europe is European. It's a well known, uncontroversial fact. It's somewhat ironic that you've asked what definitions I've used since there are wikilinks to the appropriate articles right next to the figures presented. I've just confirmed them for your benefit. I really don't see why you should have an issue with this.

Did you include Turks or not? Don't sidestep the question. It IS original research to claim that 50% of Australians are part of a European ethnic group, as no source claims that. These people are not 'from Europe', many of the people who identified with a European ancestry on the census have Australian ancestors going back to the early 1900s or even to the 1800s (I know this for a fact, as I know many of these people).

The information in the source document has been summarised into European, Asian etc because, as I've already stated, the infobox is a summary of what's in the article. Detailed lists shouldn't be included. When we did include more detail, there was an issue with people who couldn't understand that there are more Germans in Australia than Greeks now.

Well, the way the data is presented now is even more confusing except to Anglo-Celtic supremacists.

I suggest that you click on the "ANCESTRY" link in the source. If you do, you'll see that the ABS believes that "a person’s ancestry, when used in conjunction with the person’s country of birth and whether the person’s parents were born in Australia or overseas, provides a good indication of the ethnic background of first and second generation Australians".

Giving a strong indication of their ethnic background does not DETERMINE their ethnic background.

4. You are clearly missing the point. The first and last columns are all that are relevant. The first column identifies the respondent's ethnicity, the last identifies the number of respondents who identify with that ethnicity. Where the respondent's parents were born is not important. What is important is the ethnic group with which the respondents identify.

Yes, and none of them identify as European (well, 9,000 do). You obviously just lumped overlapping categories together.

::5. Because this is an article about Australia, so Australian seems the most relevant ethnicity. The source document does the same. If this was an article about India, I'd put Indian first. It's not rocket science.

Not many Indians would identify as ethnic Indians on their census. Of course, the Indian census only has caste and tribal identification but if it did have ethnic identification "Indian" would be as rare as those Yugoslavs who identified their nationality (the term used in Yugoslavia for ethnicity at the time) as"Yugoslav". So, maybe 2% of the population. Anyway, it has to be either in order of immigration to Australia (in which case Aborigines would be first) or in order of size (in which case Europeans would be first) otherwise it's just confusing (and you seem to dislike a previous table that confused people who were unaware of the fact that there was major German immigration here in the early 1900s).Saimdusan Talk|Contribs 10:48, 10 April 2009 (UTC)

"Since you haven't answered any of my questions" - Clearly, you haven't been reading my responses because I have answered your questions.

"what is your definition of European" - responded to here, 32 minutes after you asked.

"Does 'European' include Turks? Why/why not? Aren't some Turks 'from' Asia?" (etc) - There was really no need to respond to this since you answered your own question. Obviously, since it's not possible to determine which Turks were from Asia and which were from Europe, they've naturally been included with "others". You could have answered this with a calculator.

"Why not Eurasian, or for that matter, Afro-Eurasian? Why not western/northern/eastern/southern European?" - Also responded to here, as well as several other places before and after you asked the question. The information provided is a s-u-m-m-a-r-y. If I was to include terms like Eurasian, Afro-Eurasian or Anglo-Australian assimilationist ethnic nationalists, I might as well have left the data as was presented, which is not appropriate for an infobox. The continental boundaries provide a far more simple method of grouping people which is appropriate for an infobox where everything is summarised.

"The cited data shows responses to the question "What is your ancestry?". The data does not mention "ethnicity"." - responded to here, 75 minutes after you asked the question.

"Where did you get the data for 'European' anyway?" - See response to "what is your definition of European"

"You do realize that one can select two ancestries on the census, right?" - responded to here, 75 minutes after you asked the question.

Responses to new statements & questions

"Anglo - Of or relating to England or Britain." (etc) - So you created the term yourself.

"Oh come on, "whiteness" is one of the main factors of European ethnic identity." - . That's a rather racist attitude to take. Surely the main, and really the only factor, is location, hence the name.

"Did you include Turks or not?" - This is a question that you shouldn't need to ask. If I had included 56,000 turks the percentages would have been different, something easily verified with a basic calculator.

"It IS original research to claim that 50% of Australians are part of a European ethnic group, as no source claims that." - The cited source document provided exact figures for each of the ethnicities chosen by respondents. The percentage is a summary of data extracted straight from that data. It's not OR.

"many of the people who identified with a European ancestry on the census have Australian ancestors going back to the early 1900s or even to the 1800s " -

"I know this for a fact, as I know many of these people" - Now that is classic WP:OR. It's specifically mentioned in the policy, under "Citing oneself". How many of these almost 13 million people do you know and can you provide a citation from a reliable source to confirm your claim. Don't sidestep this. I want proof.

"Well, the way the data is presented now is even more confusing except to Anglo-Celtic supremacists." - More racist statements. It'd be quite ridiculous if it wasn't for the racist tone.

"Giving a strong indication of their ethnic background does not DETERMINE their ethnic background." - Why not?

"Yes, and none of them identify as European (well, 9,000 do)." - You keep stating this figure of 9,000 but so far you've failed to cite the source other than by vague statements such as "I've looked at the data" and "I got the 9,000 figure from the data itself". When are you going to provide a source for this mythical data?

"You obviously just lumped overlapping categories together" - No, I didn't obviously do that. Each identified ethnicity has been included only once.

"Not many Indians would identify as ethnic Indians on their census." - The fact is, you don't know how anyone other than you identified.

"...as rare as those Yugoslavs who identified their nationality" - Hardly relevant as Yugoslav isn't mentioned in the source data.

"Anyway, it has to be either in order of immigration to Australia " - Why? This is an article about Australia so it makes more sense to first identify those who consider themselves Australian since that is what the other ethnicities will be compared against in the first place.

"you seem to dislike a previous table that confused people who were unaware of the fact that there was major German immigration here in the early 1900s" - Perhaps you should do a little more research before commenting. What actually happened was that somebody changed the order of ethnicities.[28] Somebody else too exception to that.[29] The someone else changed Greek to German.[30] I thought that was strange, considering that Greek has always been a significant ethnicity in Australia, but, not being one to act on gut instinct, unlike some,[31] I decided to do some research. Finding the existing citation woefully inaccurate and out of date, found a better reference and updated the article accordingly,[32], listing the top 5 ethnic groups. Over subsequent weeks, "Greek" kept being added back in.[33] I reverted the first one, suggesting we keep it to the top 5 groups,[34] but Greek kept being re-added. I eventually asked myself if it was OK for the top 5 European ethnicities, why isn't it OK for the 6th, 7th etc and why shouldn't the same be the case for Asian, or any ethnicities. Unfortunately, this would blow out the size of the infobox, which is undesirable so I decided the best and fairest option would be to remove the top 5 European ethnicities altogether. Guess what? It worked. It had nothing to do with Germans and skin colour certainly didn't come into it. Sorry to disappoint you. --AussieLegend (talk) 13:54, 10 April 2009 (UTC)

Oh yes, a simple calculation of percentages. Now I believe you. AKAF (talk) 14:42, 10 April 2009 (UTC)

As you should. (a+b+c+d+e+f+g+h+i+j+k)/l isn't that hard to do, even on a calculator. A primary school kid should have no problems. --AussieLegend (talk) 14:50, 10 April 2009 (UTC)

Fine, then perhaps you will reference your source for considering Russians to be European, but Turks not to be. You appear to resist all efforts to explain that what you have done is distill a complex synthesis based on data of dubious value into a single number. Look, original research is a problem, because it defeats the purpose of correct referencing. You need to quote only exactly what your secondary sources say. If you can find a secondary source which analysis your primary source to say what you want then you're golden. Just quote the secondary source. As it is, you are being the secondary source by quoting something which is not exactly stated in the primary source. You can see this just by noting how many ways there are of calculating that percentage in the thread above. You need to just deal with the fact that the source doesn't say what you want and move on. AKAF (talk) 18:11, 10 April 2009 (UTC)

"perhaps you will reference your source for considering Russians to be European" - I'm happy to concede that Russians being included in European is an error. Thanks for picking it up. I've now fixed the error. It made a 0.26% error, although if I'd stuck to single digit accuracy as was suggested it would only be 0.2%. Funny how nobody picked it up considering I posted that almost 10 days ago.

"You need to quote only exactly what your secondary sources say." - If it was in the prose of an article I'd almost agree with you but we can't be exact in the infobox because it takes up far too much space, which is why a summary in an easy to read format is used. The reason I'd only almost agree with you is that we can't quote exactly what a source says. We're forced not to in order to avoid copyright violations. The summary is simply the same data presented in a different format. You need to deal with that. --AussieLegend (talk) 19:30, 10 April 2009 (UTC)

"Funny how nobody picked it up considering I posted that almost 10 days ago" - not really. Wikipedia editors are not here to check the details of your analysis - if it's complex enough to have a significant risk of error (which apparently it is) then it shouldn't be here at all.

In a lot of cases, copyright would require us to paraphrase sources, but this isn't one of them, because ABS content is published under the CC-BY license. So this is a non-issue (see WP:CFAQ).

The fact that the source can't be fit into the infobox without a great deal of shoehorning isn't an argument for doing that shoehorning, it's a sign that the source is not an appropriate way to fill the infobox. --GenericBob (talk) 00:38, 11 April 2009 (UTC)

"Wikipedia editors are not here to check the details of your analysis " - That's not a convincing argument because that's exactly what people, including you, have been doing, at least until it comes to responding to what I've written here.

"if it's complex enough to have a significant risk of error (which apparently it is)" - No it isn't apparent because the error was exceptionally minor (0.2%).

"The fact that the source can't be fit into the infobox without a great deal of shoehorning isn't an argument for doing that shoehorning, it's a sign that the source is not an appropriate way to fill the infobox" - Again, not a convincing argument. As you should well know, and not only because I've mentioned it several times, infoboxes are only summaries. There's very little source information that fits in an infobox without being summarised down to only a few words.

Yup, I've heard you say it repeatedly - and if it was an accurate summary that stuck to things that could be verified from the source, I wouldn't object. It's not, and lack of space is not a good enough reason for an inaccurate summary. --GenericBob (talk) 14:26, 14 April 2009 (UTC)

Note also that the presentation of the data indicates a precision of 0.01%; in that context, an error of 0.2% is not minor. FWIW, this false precision is also a problem with the analysis. The ABS makes it clear that there are all sorts of reasons why the counts published are unlikely to be exact - e.g. nonresponse, problems in coding answers, etc etc - and it's generally up to the secondary source to estimate how badly those impact on derived data. This sort of issue is one of the reasons why WP:RS warns about primary sources. --GenericBob (talk) 22:18, 15 April 2009 (UTC)

So you created the term yourself.

The term "Anglo"? Surely you've heard this term before!

Why not?

Read your own bloody source. It does not say "it determines their ethnicity". It says it gives a "GOOD INDICATION".

The fact is, you don't know how anyone other than you identified.

Um... I'm talking about the Indian census at the moment. Indians would definitely identify as Indians in the Australian census. In India however,each group is heavily divided. No, I'm not suggesting to put this into the article so it doesn't need a citation. I don't think it makes sense for you to deny such the obvious fact of India's diversity.

Hardly relevant as Yugoslav isn't mentioned in the source data.

This time, I was talking about the Yugoslav census. Remember,the article about India? I'm saying that people that identified their ethnicity as "Indian" would be a small minority, and definitely shouldn't go on the top (whatever the plurality group is, probably Bengalis as Bengali is the largest Indian language other than Hindi). You're the one that used the example of the India page, so I don't see how it's "hardly relevant".

I'm happy to concede that Russians being included in European is an error.

What about Greeks? There are Greeks native to Asia, you know. Ever heard of Greek Cypriots and Pontic Greeks?

That's a rather racist attitude to take. Surely the main, and really the only factor, is location, hence the name.

How is that racist? It's like saying that Afro-diasporic people don't have a concept of 'blackness', or that Islam isn't the largest religion of the Punjabi people. It's simply an observation about the IDENTITY of 'indigenous' Europeans and the European diaspora when they (well, we) distinguish ourselves from groups that tend to be darker-skinned and European mongrels such as myself.

One of the most important things with your data is that the numbers for 'Europeans' and 'Asians' CAN'T be added together as hundreds of thousands of Australians identified with more than one ancestry on the census.

I'm still having trouble seeing where you got your definitions from, either. If you define Asian as 'people from Asia', then we're fine, but this is about ethnic groups. Turks and Lebanese are definitely not seen as 'ethnic Asians' in Australia, as with the US 'Asian' refers to East and Southeast Asians. Similarly, Turks and Lebanese are generally not seen as 'Whites' (essentially meaning fair-skinned Europeans in this country, which is why it IS a factor of European ethnic identity) either. In fact, Greeks and Italians are only 'sometimes' seen as Whites, often being distinguished from the 'Aussies' (really, Anglo-Celtic people) as being 'Wogs'. None of this is mentioned in your source, nor does it say 'Asian', 'European' or anything else even ONCE. Saimdusan Talk|Contribs 00:22, 11 April 2009 (UTC)

"The term "Anglo"? Surely you've heard this term before!" - Of course I've heard of anglo before. As you well know, the term at issue was "Anglo-Australian assimilationist ethnic nationalists". You're just being silly with responses like that.

"It says it gives a "GOOD INDICATION"." - So, what's wrong with that? The people who responded identified the ethnic group they most identify with.

"Um... I'm talking about the Indian census at the moment." - Um... Why? The Indian census has absolutely no relevance to this issue.

"This time, I was talking about the Yugoslav census." - Again, why? How is this relevant to this or the Indian article?

"You're the one that used the example of the India page, so I don't see how it's "hardly relevant"." - I raised the Indian page only in response to your question regarding why Australian was at the top of the list. OK let's forget India. Blow it off the map. Use New Zealand instead. (Finally, a use for New Zealand!)

"What about Greeks? There are Greeks native to Asia, you know. Ever heard of Greek Cypriots and Pontic Greeks?" - We can't make assumptions that this or that response refers to Greek Cypriots, Pontic Greeks, Babylonian Italians or Martian New Zealanders. If a response says Greek then it's in the Greek box and Greece is a European country. To make any assumptions otherwise is original research. Look at the responses for "Australian Aboriginal". 1,455 responses claim that bothe parents were born overseas. How many Aboriginals have both parents born overseas? To assume that these people aren't really Aboriginal is original research. We have to include them in the Australian Aboriginal group whether we agree with it or not.

"How is that racist? It's like saying that Afro-diasporic people don't have a concept of 'blackness'," - They might have a concept of blackness but blackness isn't an an indicator that they're Afro-disaporic any more than whiteness is an indicator that somebody is European. There are plenty of Europeans of all colours.

"One of the most important things with your data is that the numbers for 'Europeans' and 'Asians' CAN'T be added together as hundreds of thousands of Australians identified with more than one ancestry on the census." - Again, this goes back to the Aboriginal example. If people identify with a particular group in the source data then we have to include them in that group or else you're resorting to original research.

"If you define Asian as 'people from Asia', then we're fine," - We're fine then.

" Turks and Lebanese are definitely not seen as 'ethnic Asians' in Australia," - As already indicated, Turks are included in others. So are Lebanes. It seems we agree.

"Similarly, Turks and Lebanese are generally not seen as 'Whites' (essentially meaning fair-skinned Europeans in this country, which is why it IS a factor of European ethnic identity)" - Again, I have to disagree with your fixation on skin colour. People from the US are generally white and they're not European. --AussieLegend (talk) 07:15, 14 April 2009 (UTC)

Australia a Continent???

This aritcle states that Australia is a continent. Forgive me for saying this, but there are only 7 continents in the world which consist of Asia, Africa, Europe, North America, South America, Antarctica, and Oceania. Therefore, Australia is not a continent and stating such in this article is false. Australia is nation within Oceania. --Yoganate79 (talk) 02:40, 8 March 2009 (UTC)

Please see the first sentence of the Continent article and its supporting citation. Nick-D (talk) 02:44, 8 March 2009 (UTC)

As well as the numerous discussions about this very issue in the Talk page archives. Oceania is not a continent. It is a region. --AussieLegend (talk) 06:47, 8 March 2009 (UTC)

Australia is the continent in the list you give, not Oceania, most of which is small island nations. --Dmol (talk) 08:01, 8 March 2009 (UTC)

Having been involved in s similar discussion on this page, I'd like to suggest to Yoganate that he leaves it alone. "Continent" is a poorly defined term, which you can pretty much let mean what you want. So just let those editors continent themselves with their inaccuracy. Australians have been taught in school that "Australia is the largest island and the smallest continent", and pretty much everyone who edits this page believes that to be an important part of the red white and blue running through their veins. Leave alone that a goodly amount of what gets taught in school is bullshit: "Australia is the only place where marsupials are found" or "Australia was one of the first countries to have universal suffrage" occur immediately as examples. It appears to bother these editors not one whit that this argument leaves Tasmania, New Guinea, and most of Indonesia unattached to any continent. Moreover it appears to bother noone that this entire set of statements predates continental plate theory, Or that this statement was made as the protectorate of New Guinea was a part of Australia. One might as well write the whole article from the point of view of 1960s historians, with the noble attempts at integration of the poor abo kids into generous white families. However seeing as the statement about continents is absolutely unbased in any reality, and its just a throwaway line, its probably not worth it to argue over such a triviality. Except that its location in the first paragraph it gives the reader the impression that the scholarship for the whole article will be as slipshod. AKAF (talk) 15:10, 15 March 2009 (UTC)

I remember you. You were the editor who claimed that the Australian Government's Department of Foreign Affairs and Trade website was an advertsing website and not a reliable source.[35] Welcome back and please remember, should you choose to present any arguments, rather than just attack the editors of this article and Australians in general, some civility will result in your comments gaining a better response than the incivility that you've just demonstrated. --AussieLegend (talk) 16:06, 15 March 2009 (UTC)

Here's the truth about the government: they have departments whose purpose is to promote some aspect of Australia to the country and overseas (ie "Advertising", for some value of that word). And yes, the department of foreign affairs and trade is not a reliable reference for the same reason that an autobiography is not the authoritative guide to a person's life. In fact most government departments are not good references for an article about that department's government, and should be regarded as a primary source for the purposes of citing. So you'll excuse me if I find referencing such a site to be at best disingenuous. AKAF (talk) 08:18, 16 March 2009 (UTC)

Continent discussions - for future reference

Past "Is Australia a continent?" and "Australia is not a continent" discussions, and related topics, may be found at:

• Talk:Australia/Archive 3#In Australasia or only country in a continent
• Talk:Australia/Archive 3#Excessive categories
• Talk:Australia/Archive 4#The whole continent?
• Talk:Australia/Archive 5#Australia or its commonwealth
• Talk:Australia/Archive 6#Island
• Talk:Australia/Archive 8#Landmass
• Talk:Australia/Archive 13#Australia is not a continent
• Talk:Australia/Archive 14#Australia is the only continent occupied by a single country
• Talk:Australia/Archive 14#Is Australia a country or a continent?
• Talk:Australia/Archive 14#Australian Continent

I could have sworn it's been discussed more than that. --AussieLegend (talk) 09:15, 8 March 2009 (UTC)

It really is quite sad that Australians are brainwashed and uneducated to think that their country is one of the classical continents found on earth. My question is, if Australia is considered a continent, then what continent does New Zealand belong to then? In any case, Greenland should also be classified as a continent then too. Let's classify Long Island, Cuba, Japan, and the British Isles their own continents too. --Yoganate79 (talk) 01:52, 26 March 2009 (UTC)

Excuse me? Brainwashed and uneducated? How insulting! Bidgee (talk) 01:57, 26 March 2009 (UTC)

Exactly australia has better education systems than america. At least we can go to school without getting shot.

Australia a continent yes but australia is the figurehead of australasia a.k.a. oceania new zealand, papua new guinea and that fall under that continent.— Preceding unsigned comment added by 124.171.106.175 (talk • contribs)

Australasia is not "a.k.a." Oceania, it's a region of Oceania. Neither are continents. --AussieLegend (talk) 09:16, 9 April 2009 (UTC)

Countries do not have to belong to a continent. Sorting out which continent the Maldives is located would be tricky. Furthermore, Australia is not a third world country with a tyrannical government. It terms of public brainwashing and lack of education, I'd encourage you to visit someday, and see if your comment is justified. ∗ \ / (⁂) 11:46, 9 April 2009 (UTC)

remains a Commonwealth realm

Do Editors think it is appropriate to say in the opening para that Australia "remains a Commonwealth realm"? Shouldn't the article say Australia is a "constitutional monarchy" - as that term is used in the UK article and Canada's. This is being discussed on Talk:United Kingdom. Regards. Redking7 (talk) 11:48, 15 March 2009 (UTC)

I support using the "Constitutional monarchy" description in the opening para:

1. Redking7 (talk) 11:48, 15 March 2009 (UTC)
2. Hawkeye7 (talk) 03:25, 18 March 2009 (UTC)
3. Slac speak up! 04:52, 25 March 2009 (UTC)
4. --Michael Johnson (talk) 07:30, 30 March 2009 (UTC)
5. Night w (talk) 06:13, 22 April 2009 (UTC) (P.S. I also don't think it's a big deal).
6. [Insert your username here]

I support using the Commonwealth realm" description in the opening para

1. [Insert your username here]

Comments/Questions I need more detail in why should it be changed, whats more widely used (Commonwealth realm or Constitutional monarchy) in Australia (not just within the Commonwealth). Bidgee (talk) 11:52, 15 March 2009 (UTC)

I'd also like to know why it should be changed. I should point out that Wikipedia works on consensus, not voting, so the polls here and at Talk:United Kingdom are premature. It's not appropriate to make a post on one article's talk page and declare consensus at another, which is what prompted Bidgee's reversion.[36] You need to actually discuss the issue before consensus can be reached and this particular matter needs to be discussed here. --AussieLegend (talk) 12:26, 15 March 2009 (UTC)

Have a look at Talk:United Kingdom - its being discussed there too (more editors are involved on that page). Regards. Redking7 (talk) 21:59, 16 March 2009 (UTC)

The discussion at Talk:United Kingdom is about use at United Kingdom, not here. As I indicated on your talk page,[37] any consensus achieved at Talk:United Kingdom does not immediately become applicable here. You need to discuss the issue here, not there and explain why the change should occur here. --AussieLegend (talk) 01:19, 17 March 2009 (UTC)

I support the reference. However, the purpose of it is to get across that when the colonies federated in 1901, the new country did not "go it alone" as a republic, but continued to be part of the Commonwealth (or Empire as it was then) and continued to acknowledge the British monarch as its monarch. It's also a constitutional monarchy of course, but maybe the purpose of this reference to Commonwealth realms could be spelled out a little more explicitly. -- JackofOz (talk) 23:52, 16 March 2009 (UTC)

Constitutional monarchy is a widely understood term. Commonwealth realm is a term not used in Australia, is not widely understood, and is not as descriptive as constitutional monarchy. Hawkeye7 (talk) 03:30, 18 March 2009 (UTC)

Yes, but just saying it remained/became a constitutional monarchy could mean that it set up its own monarchy, distinct from the British monarchy. We need to communicate the continuity of Australia's allegiance to the British monarchy between pre-1901 and post-1901. -- JackofOz (talk) 07:51, 18 March 2009 (UTC)

What do the sources say? I'm and Australian, and I can't say that I've ever seen a description of Australia as being a 'Commonwealth realm'. A Google search of 'Australia Commonwealth realm' doesn't produce any reliable-looking sources in the first couple of pages of results ([38]), and the Queen's website states that "Australia is a constitutional monarchy with The Queen as Sovereign": [39], though Australia is in the drop down list of 'Commonwealth realms' at the top of the page. Nick-D (talk) 07:59, 18 March 2009 (UTC)

Well I've found this (Paper crown when Charles takes reign) which seems to point that Australia is a Commonwealth realm but again I request for more detailed information with sources before I can make a decision. Bidgee (talk) 08:15, 18 March 2009 (UTC)

Why can't both have a place in the lead? Here is a suggestion:

“ For around 40,000 years before European settlement commenced in the late 18th century, the Australian mainland and Tasmania were inhabited by around 250 individual nations of indigenous Australians. After sporadic visits by fishermen from the immediate north, and European discovery by Dutch explorers in 1606, the eastern half of Australia was claimed by the British in 1770 and initially settled through penal transportation to the colony of New South Wales, founded on 26 January 1788. The population grew steadily in the following years; the continent was explored, and during the 19th century another five largely self-governing Crown Colonies were established. On 1 January 1901, the six colonies became a federation, and the Commonwealth of Australia was formed.

The Commonweath is a constitutional monarchy with strong liberal democratic traditions. As a Commonwealth realm, Australia's monarch is Elizabeth II. The population is just over 21.7 million, with approximately 60% concentrated in and around the mainland state capitals of Sydney, Melbourne, Brisbane, Perth, and Adelaide. The nation's capital city is Canberra, located in the Australian Capital Territory (ACT).

”

It is just one possibility. -Rrius (talk) 03:29, 26 March 2009 (UTC)

I agree, I don't see why it should be a case of one or the other. I think your suggestion is an excellent way of incorporating both truths. --AussieLegend (talk) 03:44, 26 March 2009 (UTC)

For the information of RedKing, "Commonwealth Realm" is a *very* infrequently used term in Australia, even in constitutional law circles (my constitutional law textbook didn't mention it once). "Constitutional monarchy" is vastly more common; I would even go so far to say that this is an imported Canadism, like "Her Majesty's Loyal Opposition" (never used in any official context in Australia). Slac speak up! 02:27, 30 March 2009 (UTC)

We are, nevertheless, both things. We're a constitutional monarchy because we're governed by a constitution and we have a monarch (regardless of who that monarch is); but because the monarch happens to be Elizabeth II, we're also a Commonwealth realm. Not all constitutional monarchies are Commonwealth realms, but all Commonwealth realms are constitutional monarchies. -- JackofOz (talk) 05:13, 30 March 2009 (UTC)

Well Australia is many things, but we don't have to include them all in the opening paragraph. Commonwealth realm may be an accurate description, but it is not one that strikes me as being commonly used in relation to Australia. I'm not fussed, but I can't see the point of including it. --Michael Johnson (talk) 07:30, 30 March 2009 (UTC)

It may not be particularly often used in Australia, but if we write the article using only terms that are frequently used in Australia, it would be of less value to international readers. There seems little point in having an article called Commonwealth realm without mentioning - not just in that article but also in this one - that Australia is such an animal. -- JackofOz (talk) 12:06, 30 March 2009 (UTC)

Can't it just say ...maintained a stable liberal democracy and a constitutional monarchy within the Commonwealth. (or something similar and better phrased)? Why not use both terms and get rid of that vague "political system" term? Night w (talk) 06:20, 22 April 2009 (UTC)

citation needed

The article is currently a FA, but how come it still has [citation needed] templates in a few sections? Aaroncrick^{(Tassie Boy talk)} 09:04, 16 March 2009 (UTC)

Because it is a high visibility article that is edited by a hell of a lot of people, making it difficult to maintain at FA standard for any length of time. This was featured in June 2005. It probably doesn't deserve to be featured now. Hesperian 01:22, 17 March 2009 (UTC)

The FA criteria probably wasn't as strict back then. Although I must admit I was a bit surprised when i saw that is was a FA. Aaroncrick^{(Tassie Boy talk)} 09:41, 18 March 2009 (UTC)

What about removing the info? That normally sparks a reaction - hopefully to find the cite. --Merbabu (talk) 23:01, 20 March 2009 (UTC)

Well there's a thought! Never thought about it that way. Aaroncrick^{(Tassie Boy talk)} 09:56, 7 April 2009 (UTC)

Category "Liberal democracies"

I think there is no more absolute consensus that Australia is a liberal democracy beacuse of strict Internet censorship. Therefore I will remove this article from category Category:Liberal democracies. —ilaiho (talk) 17:53, 22 March 2009 (UTC)

User:AussieLegend undid my edit, saying there is only one censored website. According to some recent news and that Wikipedia article, I have understood that there is much more irresponsible censorship than "one censored website". Anyway, this categorization is not very important for me. —ilaiho (talk) 17:07, 24 March 2009 (UTC)

The United States, United Kingdom and New Zealand all censor the internet, and remain on the category Category:Liberal democracies. Hawkeye7 (talk) 22:05, 24 March 2009 (UTC)

The idea that the government's censorship regime disqualifies Australia as a liberal democracy is absurd (as much as I may detest its planned internet filter scheme). Plenty of liberal democracies have more stringent censorship than Scandinavian countries (in fact, most do). Slac speak up! 04:54, 25 March 2009 (UTC)

Firstly, the ludicrous internet filtering proposal is currently a trial and is not (yet) an actual reality. Secondly, as the article on Liberal democracy points out, all such political structures do contain specific limits on specific freedoms. Every liberal democracy contains some internet restrictions on (for example) child pornography. As the article on the topic makes clear, the core of a liberal democracy is not any particular social law but the citizens elective rights, including free and fair elections and a competitive political system. Euryalus (talk) 09:47, 7 April 2009 (UTC)

How can a country with the Queen as a state head be a democracy??? And liberal??? Sorry but it doesn't work together. Monarchy is not a democracy. —Preceding unsigned comment added by 61.69.27.220 (talk) 06:22, 14 April 2009 (UTC)

Demography

Under the demography section, the information that only 70 aboriginal languages still exist is incorrect. According to the 2005 Ethnologue, which is the most authoritative linguistics survey, there are 280 recorded aboriginal languages, 238 still spoken and 42 extinct. I don't know how to change this since the article is protected. Also, there is no source for the claim that an indigenous language is the main language for 50,000 people. Gordon, Raymond G., Jr. (ed.), 2005. Ethnologue: Languages of the World, Fifteenth edition. Dallas, Tex.: SIL International. Online version: http://www.ethnologue.com/. Claritarejoice (talk) 17:55, 23 March 2009 (UTC) Claritarejoice

Australia Interactive Map

I suggest the Link below containing Google Interactive Map of Australia . http://www.all-maps.info/2009/03/australia-map.html —Preceding unsigned comment added by Kewaga (talk • contribs) 14:20, 30 March 2009 (UTC)

Business Address?

Can you please add Australia's business address to this article. It is 1601 MASSACHUSETTS AVE NW C/O AUSTRALIAN EMBASSY WASHINGTON DC 20036 and Australis's company number is 0000805157. Source - http://www.sec.gov/cgi-bin/browse-idea?CIK=0000805157&action=getcompany

Umm, no, I don't think so. --Michael Johnson (talk) 01:30, 3 April 2009 (UTC)

Please keep in mind that Australia is also well known as the ' Down Under ' and the ' Wide Brown ' . Lejon (talk) 04:49, 6 April 2009 (UTC)

Never heard of Wide Brown. Aaroncrick^{(Tassie Boy talk)} 09:57, 7 April 2009 (UTC)

What are they teaching kids at school these days?

I love a sunburnt country,

A land of sweeping plains,

Of ragged mountain ranges,

Of drought and flooding rains.

I love her far horizons,

I love her jewel-sea,

Her beauty and her terror

The wide brown land for me!

If you say you've never heard of that..... --AussieLegend (talk) 10:31, 7 April 2009 (UTC)

I think it's a good idea, but maybe we should post it to America's business address. Um, what's the US' business address? The US business address is being moved to Shanghai, right?--Merbabu (talk) 10:35, 7 April 2009 (UTC)

Nah, it's somewhere in Calcutta. -Rrius (talk) 12:34, 7 April 2009 (UTC)

Australia?Austria?

Maybe we should put a

Not to be confused with Austria.

here.People often make such mistake.Especially non-English native user like me.囧--半弯不直男 (talk) 13:54, 8 April 2009 (UTC)

RfC: Is data in infobox 'ethnic groups' field adequately supported by source?

A user has requested mediation on this issue. A mediator will be here shortly to assist you. The case page for this mediation is located here.

Template:RFCsoc

Summary: Original source[40] presents results of a Census question in which respondents were allowed to give up to two responses to the question "What is the person's ancestry?" Source gave (a) total number of responses for each of several fine-level categories (e.g. "Scottish", "Italian", and "Ancestry not stated"), (b) total responses (approx. 25 million), and (c) total persons (approx. 20 million - discrepancy due to people providing more than one response.)

Infobox on Australia uses this data in the 'ethnic groups' field. Categories have been converted to broader supercategories not explicitly named in the source (e.g. "European", "Asian") and data are expressed as percentages rather than totals. Percentages are calculated as the ratio of responses in relevant categories/supercategories to total responses.

Some points of contention are:

- Is it original research to represent data compiled and published on the basis of 'ancestry' as 'ethnic groups'?

- Is it OR to categorise individual countries into broader entities? (e.g. deciding whether to count Turkey as part of 'Europe'?)

- Is generating totals for supercategories by summing fine-level categories a routine calculation, or does the complexity of the data (specifically, the potential for a single person to be counted in more than one fine-level category due to multiple responses) make this OR?

- Is converting totals to percentages a routine calculation, or does the complexity of the data (specifically, the question of whether to divide by total responses or total persons) make this OR?— Preceding unsigned comment added by GenericBob (talk • contribs)

I suggest that you read the "Quality Statement - Ancestry" section of this ABS article (linked to from within the source document), which states "The ancestry data collected in the Census is a measure of self-identification of ethnic or cultural group affiliation and therefore provides a broad measure of cultural diversity."" There's the link to ethnicity.

Another interesting statement is "Scottish ancestry (in terms of responses for both ANC1P and ANC2P) has risen from 2.4% of persons in 2001 to 5.9% of persons in 2006"." It's interesting because if you divide the number of responses reporting Scottish ancestry by the number of total responses (534,882/22,568,471 and 1,501,201/25,410,601 for 2001 and 2006 respectively) you arrive at the results of 2.4% and 5.9%, the figures quoted in the ABS document. On the other hand, if you divide by total persons, as GenericBob would have us do, the percentages arrived at are 2.9% and 7.6%, which are significantly different to the figures quoted by the ABS. The method used to arrive at the figures in the infobox is the same used by the ABS, i.e. responses divided by total responses. --AussieLegend (talk) 15:16, 10 April 2009 (UTC)

Regarding the issue of counting Turkey as part of Europe, this was not done so I don't see why it is a point of contention. As I have indicated, and as can be confirmed by converting the percentages back to raw data, Turkey has been included in "Others" as the Turkish responses can not be reasonably included in either Europe or Asia without making assumptions. --AussieLegend (talk) 15:24, 10 April 2009 (UTC)

It is important for wikipedians to understand proper academic citation and the value of excluding material which is nice for the narrative developed for the reader, but which is mightily truthy. If you reference something from a source, you can't analyse the data from the source to get your result. You can look at a secondary source which does that analysis, but wikipedia is not about original analysis, or WP:OR in the local jargon. It is pretty obvious from the comment above that analysis is happening. I find it unlikely that any of the participants on this page are qualified demographers, and if so, it wouldn't matter, since there are other places for demographers to publish their original research. AKAF (talk) 18:25, 10 April 2009 (UTC)

You are being ridiculous in claiming that analysis is being done. The only "analysis" being done is to group ethnicities by continent, which requires absolutely no analysis. Is England in Europe? Is Germany in Europe? Is Japan in Asia? Is Vietnam in Asia? These are the only questions being asked and the answers require no analysis. --AussieLegend (talk) 19:03, 10 April 2009 (UTC)

So, is Russia in Europe? The fact that you had to reconsider this issue should be a pretty strong clue that this is nontrivial analysis.

I agree that the quality statement certainly supports the idea that ethnic group affiliation is related to ancestry data, I believe it's OR to carry this to the point of treating them as the same thing - note that "ethnic or cultural group" in there. As I noted earlier, the ABS documentation says that ancestry is a "good indication" of ethnic background of first and second generation Australians, when used in conjunction with the person's country of birth.

Do you believe the data we're discussing here pertains only to first- and second-generation Australians? Do you believe that it's being used in conjunction with the person's country of birth? (Hint: country of birth is a separate data item, collected in a different question, and none of the data from that question is included in the source document you're citing.)

Scots issue - I'm going to delay my answer to that until after the public holiday, since it will require giving somebody a phone callcontacting someone during business hours. --GenericBob (talk) 23:37, 10 April 2009 (UTC)

OK, I've discussed this issue with a staff member in the ABS' Census Products & Services section, noting that the verbal description of the Scots figures ("x% of persons") is inconsistent with the data (apparently calculated as a percentage of responses). Her response was that "it would be far more sensible to calculate based on persons rather than responses", and after checking with those responsible for the quality statement, it will be amended from this:

"Scottish ancestry (in terms of responses for both ANC1P and ANC2P) has risen from 2.4% of persons in 2001 to 5.9% of persons in 2006 and this may be partly due to its inclusion in the list of check box responses..."

to this:

"Scottish ancestry (in terms of responses for both ANC1P and ANC2P) has risen from 2.9% of persons (2.4% of all responses) in 2001 to 7.6% of persons (5.9% of all responses) in 2006. This may be partly due to its inclusion in the list of check box responses..."

Not sure how long it will take to get a corrected version on the website, but in the meantime, if you're not content to take an anonymous Wikipedia editor's word for it ;-) you can verify this by contacting the ABS. Alternately, while I'd rather not give out my real name here, I'd be happy to provide a copy of the email exchange to one of the Aussie admins. Edit 20/4/09: This correction has now appeared on the ABS website, worded as above.

Note that this quality statement was never presented as a statement on how percentages should be calculated - the purpose of this section was to explain a different issue (i.e. the fact that adding a checkbox for 'Scottish' had greatly increased the number of such responses - something that's apparent regardless of whether you divide by total responses or persons). This is probably how the error managed to slip through, and it's a reminder of why we should stick to things that are explicitly stated in the source rather than attempting to extrapolate from sources that weren't written - or checked - with that particular extrapolation in mind. --GenericBob (talk) 05:13, 14 April 2009 (UTC)

Note - since it makes more sense to respond to the following paragraph by paragraph, I've broken up AussieLegend's comments and added his signature to each paragraph to keep attribution clear. If this is a problem, please feel free to restore the original structure and move my responses to a single section at the bottom. --GenericBob (talk) 07:33, 14 April 2009 (UTC)

"The fact that you had to reconsider this issue" - I didn't reconsider anything. A mistake was identified and I immediately corrected it because I had obviously just placed Russia in the wrong pile. All the calculations were done on a spreadsheet so it's an easy thing to do. I note that you never challenged it so, using your logic, you must have agreed that Russia is in Europe.

"Do you believe the data we're discussing here pertains only to first- and second-generation Australians?" - The source doesn't just identify 1st and 2nd generation Australians. It deals with every Australian.

"if you're not content to take an anonymous Wikipedia editor's word for it" - Not just me, it's policy. We're not reliable sources so what you or I say means nothing unless we can back it up with a citation from a reliable source. I've cited a reliable source showing that I used the same methodology as the ABS to calculate the percentages.

"Note that this quality statement was never presented as a statement on how percentages should be calculated " - However, it does identify the methodology used by the ABS which is extremely relevant. It was clear from the Census dictionary statement that the ABS determined percentages by dividing responses by responses. In order to summarise the data supplied by the ABS it's necessary to use the same methodology in order to come up with the same figures. To not do so, eg dividing responses by persons, is original research. If the ABS do change the census dictionary then the methodology will have to change because then, and only then, will dividing responses by responses become OR. Until that time, divividing by responses is appropriate, dividing by persons is OR. --AussieLegend (talk) 05:54, 14 April 2009 (UTC)

OK, the ABS have now corrected the page, making the changes that I indicated last week - are we all now agreed that choosing 'total responses' as the appropriate measure is OR? --GenericBob (talk) 04:41, 20 April 2009 (UTC)

Supercategories

AL: "The fact that you had to reconsider this issue" - I didn't reconsider anything. A mistake was identified and I immediately corrected it because I had obviously just placed Russia in the wrong pile. All the calculations were done on a spreadsheet so it's an easy thing to do. I note that you never challenged it so, using your logic, you must have agreed that Russia is in Europe.

Er, no. I make no statement either way on whether Russian (or Turkish) ancestry counts as 'European'. Those are knotty questions, and I believe any such assessment would be OR; whether Russians and Turks are 'European' depends on the individuals and on definitions used. Silence does not imply agreement, it only implies that I have a finite amount of time to spend here. Checking the details of your calculations - when my point is that those calculations don't belong here in the first place - would not be a good use of my time. --GenericBob (talk) 07:33, 14 April 2009 (UTC)

Equation of 'ancestry' and 'ethnic group'

AL: "Do you believe the data we're discussing here pertains only to first- and second-generation Australians?" - The source doesn't just identify 1st and 2nd generation Australians. It deals with every Australian.

Agreed. So do you acknowledge that we can't justify equating 'ancestry' to 'ethnic group' by relying on a source which only draws a connection between the two in the context of first and second generation Australians and when used in conjunction with the person's country of birth? --GenericBob (talk) 07:33, 14 April 2009 (UTC)

Amendment to source material

AL:"if you're not content to take an anonymous Wikipedia editor's word for it" - Not just me, it's policy. We're not reliable sources so what you or I say means nothing unless we can back it up with a citation from a reliable source. I've cited a reliable source showing that I used the same methodology as the ABS to calculate the percentages.

You have misunderstood WP:RS - the sourcing standards given there are a necessary criterion for inclusion, not a sufficient one. It states that content included in Wikipedia should be supported by 'reliable sources', as defined therein; it does not state that everything found in such sources should be used. (That would be an 'infallible source', and at that point we're arguing religion, not demography.)

In this case, you have cited a 'reliable source' that (when taken along with the included figures) contradicts itself. The wording currently says that the figure given is a percentage "of persons", while we can deduce from the figures that it was calculated as a percentage of total responses. The fact that it contradicts itself makes it pretty clear that one of those two things is in error - either 'persons' was the correct method and the calculation was mistakenly done on total responses, or 'total responses' was the correct method and the documentation is using the wrong terminology.

It should be abundantly clear that a source that contradicts itself on a specific issue is not reliable on that particular issue, no matter how well it satisfies the letter of WP:RS. That on its own, without any clarification from ABS, is a pretty good reason not to use it for Wikipedia content. The fact that several other editors have argued against your approach should be another good reason to realise that what you're doing is nontrivial interpretation - and hence OR. If you're determined to keep arguing the point until that clarification comes out, you'll have a third good reason - but the longer you keep digging, the longer it takes to climb out afterwards. --GenericBob (talk) 07:33, 14 April 2009 (UTC)

Scope of source material

AL:"Note that this quality statement was never presented as a statement on how percentages should be calculated " - However, it does identify the methodology used by the ABS which is extremely relevant. It was clear from the Census dictionary statement that the ABS determined percentages by dividing responses by responses. In order to summarise the data supplied by the ABS it's necessary to use the same methodology in order to come up with the same figures. To not do so, eg dividing responses by persons, is original research. If the ABS do change the census dictionary then the methodology will have to change because then, and only then, will dividing responses by responses become OR. Until that time, divividing by responses is appropriate, dividing by persons is OR.

It was also clear from the Census dictionary statement that the ABS felt it was appropriate to discuss percentages 'of persons' (and I think we are agreed that, whatever they might be, the percentages currently on Australia are not percentages of persons?) I have said this several times already, but apparently I need to repeat it once more: I am not arguing for the infobox to include figures calculated by dividing by persons - even when the ABS quality statement is corrected, that would still be OR because of the other issues mentioned above (namely, equation of 'ancestry' to 'ethnic group' and formation of supercategories). I am arguing that the ABS source we're discussing cannot be used to fill this data field without committing OR. The only reason I'm mentioning alternate (and better) ways to analyse that data is to underline the fact that the sort of demography you're doing here is not as straightforward as you believe. --GenericBob (talk) 07:33, 14 April 2009 (UTC)

I'd like to take a moment and point out that it's considered poor form to respond the way you have,[41] despite the explanation. I'm assured that responding as I did here, quoting the protions of the respective comments that you wish to reply to, is preferred. Editing others comments is frown upon in general[42] and, even though you have given permission, it really shouldn't be up to me to re-edit your comments so that they're in the appropriate form. --AussieLegend (talk) 07:56, 14 April 2009 (UTC)

It's not universally verboten; for instance, as mentioned in the last of your links there, "In some cases, it is OK to interrupt a long contribution, either with a short comment (as a reply to a minor point) or with a heading (if the contribution introduces a new topic)." I felt this was a long contribution that could usefully be divided into subtopics without taking you out of context (and IMHO this discussion is painfully long already), but per your request I've re-edited to restore your original comment in full.

By the same token, perhaps you could refrain from snide edit summaries[43][44][45] in future? --GenericBob (talk) 08:50, 14 April 2009 (UTC)

(offtopic, but...) GenericBob, often people do replies as you were thinking about above, by quoting parts of the text of the other person in italics. This gets around the problem of editing other people's comments. I personally find AussieLegend's method of quoting to be confusing and counterproductive because he takes too little of the original text and so I end up spending half my time trying to decide who wrote what. I personally prefer to either quote a larger portion, or to quote nothing at all. Personal preferences vary greatly across the project. AKAF (talk) 09:16, 14 April 2009 (UTC)

You don't need to quiote everything the person has said, just the important points. Quoting everything just makes the discussion ridiculously long, as does creating numerous Level 2 headings for each point of a reply. --AussieLegend (talk) 09:38, 14 April 2009 (UTC)

Yep, I set the header levels wrong. Should be fixed now. --GenericBob (talk) 14:20, 14 April 2009 (UTC)

I never disagreed with that. I just noted that I find your style of quoting hard to read.AKAF (talk) 10:26, 14 April 2009 (UTC)

The point is, you didn't divide into sub-topics. You just broke apart a response by one editor. You haven't even divided into sub-topics now. You've broken out into individual main topics which makes it even more confusing to follow. As for the snide comments:

[18] - The same editor continued to ask the same question multiple times, even though answers to the question had been given multiple times. The edit summary wasn't snide, it was a summary of what I'd expressed in the edit.

[19] - Not snide, it was an attempt at some humour, based on what the editor had said and the way he had presented himself. You really need to be far less serious and try to laugh occasionally.

[20] - It was you who tried to introduce a different methodology for calculating percentages to that used by the ABS, creating confusion and trying to introduce original research while criticising me for doing the same thing, when all I'd done was summarised presented figures from the source using the same methodology used by the author of the source. If you're going to be critical of somebody, be critical of yourself, both for that and for not acknowledging, after I showed it to be the case, that your methodology was not the same. The edit summary expresses exactly what had happened. Your maths was bad. --AussieLegend (talk) 09:38, 14 April 2009 (UTC)

"trying to introduce original research"... nope. The only content change I have ever advocated to this section is removing the material currently in this part of the infobox. I have certainly argued that there is a better approach than the one you took, but as I've explained several times already - I'm sure other editors are getting tired of hearing me say this - I am not arguing that that better approach should be used to populate the infobox. I mention it to illustrate that this is not a trivial issue, and in deciding to go with one half of a self-contradicting source rather than the other half you are performing OR.

I have never contested the claim that the "2.4%" figure was derived by the 'total responses' method - while there is nothing on the ABS website (currently) to confirm it, I'd guess you're correct in concluding that that was the method used. What I do contest is your assumption that an undocumented method (contradicted by the documentation, at that) that could only be determined by reverse-engineering a calculation used only to discuss the effects of a particular checkbox on the form should be treated as the ABS' prescription for how figures should be calculated in another context. Doesn't that seem even a teensy weensy bit like OR?

Sometimes even the ABS website has errors. In this case, you've been unfortunate enough to latch onto the error first instead of the correct approach described in the same page - which is the sort of thing that happens to all of us once in a while. But when other editors point out that there are numerous problems with your method - and that the ABS website contradicts itself on this issue - and that you can very easily call the ABS to confirm that you're using the wrong method - trying to save face by accusing them of incompetence is an unwise tactic. As it happens, I work with this sort of stuff for a living (and I'd be happy to substantiate that to any admin willing to drop me an email address); while there are plenty of people around who can teach me things about demography, I've seen enough here to satisfy me that you're not one of them. --GenericBob (talk) 14:17, 14 April 2009 (UTC)

(unindent) Actually you still haven't answered the question I asked multiple times, so I feel that your snook is unreasonably cocked. In a tertiary source, you can't go re-analysing the data from a primary source, so you have only two possibilities:

1. Either: WP:OR is violated.
2. Or: You can show where the source says exactly what you quote.

What we "source" in wikipedia are directly attributable data, which can be traced back to verbatim quoted from the source material. So show me the verbatim quote or remove the material as WP:OR. There is no third choice. AKAF (talk) 10:35, 14 April 2009 (UTC)

• I made a change to the infobox here. cygnis insignis 10:54, 14 April 2009 (UTC)

Since we seem to be at an impasse, I'm sticking up a request for mediation. Further context on this dispute is in Talk:Australia#29_percent_of_people_in_Australia_Australian.3F above and Talk:Australia#Removal_of_Ethnic_Group below.

Removal of Ethnic Group

None of the groups in the "ethnic group" category are actually ethnic groups. Most of them are geographic designations...and the term "Australian" is a heritage term at best.

Intranetusa (talk) 04:41, 16 April 2009 (UTC)

You're going to need to do a lot better than that to justify removal. As has been explained above, the information in the infobox is a summary of ethnicities based on geographic origin. Summaries are used because information in the infobox needs to be brief. The CIA Fact Book, which is widely accepted as a reliable source, identifies Australian, Asians and Europeans in its list of ethnic groups.[46] Notable, reliable sources such as Encyclopædia Britannica interchange the terms "ancestry" and "ethnicity", even within the same document.[47] --AussieLegend (talk) 09:14, 16 April 2009 (UTC)

Removed again since the data is unsourced (see about half of this talk page). AKAF (talk) 09:35, 16 April 2009 (UTC)

The data is sourced. There's a citation attached, the methodology for summarising the data has been explained and proven and the question about ethnicities has been supported with references to two very reliable sources. There has been no consensus reached to remove this information. Ethic information has been in the article for a long lime so, until there is consensus to remove it then it shouldn't be removed again. --AussieLegend (talk) 10:28, 16 April 2009 (UTC)

But the citation is bullshit, as has been explained to you by a large number of people. However I'm at the point where you win this WP:GAME. Best of luck with it. AKAF (talk) 10:38, 16 April 2009 (UTC)

"The CIA Fact Book, which is widely accepted as a reliable source, identifies Australian, Asians and Europeans in its list of ethnic groups." No, it does NOT. The groups it lists are "white 92%, Asian 7%, aboriginal and other 1%". (Note that those percentages also conflict with the ones currently in the article that you're trying to defend - you've listed "5.17% Asian".)

"the methodology for summarising the data has been explained and proven" - assertion is not proof. I've repeatedly pointed out the fallacy of using 'ancestry' alone as 'ethnic group' when the source makes it clear that ancestry should be considered in combination with other information, and you've repeatedly ducked the issue. I've repeatedly pointed out that a source that contradicts itself on an issue (as the ABS page currently does on 'persons' vs 'total responses' method) cannot be considered reliable on that issue, and you've repeatedly ducked the issue. And so on and so on.

"until there is consensus to remove it then it shouldn't be removed again" - you might want to read WP:CCC, which specifically explains why this argument doesn't hold water. In any case, you seem to be the only person still arguing for inclusion of this material - and 'consensus' is not 'unanimity'.--GenericBob (talk) 12:17, 16 April 2009 (UTC)

"No, it does NOT" - Yes, it does SO. Look in the Ethnic Groups column for Norfolk Island where it says "descendants of the Bounty mutineers, Australian, New Zealander, Polynesian". You need to look at everything and not be so limited and selective in your focus. I'll let you find European on that page. I did. --AussieLegend (talk) 12:39, 16 April 2009 (UTC)

Yes, but to quote directly from that source [48] "Ethnic groups: white 92%, Asian 7%, aboriginal and other 1%". See, that's the difference between synthesis and proper citation, namely that I can directly quote the source, and it says what I want. Now do that for your source. Is there actually more than one person who believes that the current version is correct? AKAF (talk) 16:50, 16 April 2009 (UTC)

And based on the same page you've cited, the July 2009 (yes, 3 months from now) population estimate is 21,262,641. That's great except that, as of right now, the official Australian population clock[49] estimates the population to be 21,733,346, almost 471,000 more than the CIA's estimate. Based on the current net increase of 1 person every 90 seconds, which is what it has been for some time now, we should have around 21.8 million on 1 July, 542,000 more than the CIA estimate. Maybe the CIA knows something the ABS doesn't. Yes, that page may say what you want, but what it says is wrong. --AussieLegend (talk) 17:49, 16 April 2009 (UTC)

Ding, ding, ding. Yes Alex I know the answer! What is a WP:PRIMARY source? AKAF (talk) 18:35, 16 April 2009 (UTC)

Are you really going to argue that the CIA has a better idea of Australia's population than the ABS does? --AussieLegend (talk) 23:26, 16 April 2009 (UTC)

A bit more detail on population estimation, and I apologise for the length of this digression:

First off, there are several different ways to define the 'population' of Australia. Do you count foreign diplomatic staff? US military personnel passing through? Australian residents who are currently overseas? The answer depends on what you want to do with the data (or, sometimes, on what you can most consistently collect.)

Every five years, the ABS runs a census. This counts most of the people in Australia, with a few deliberate exclusions (e.g. foreign diplomatic staff, IIRC). A substantial number of people get missed for one reason or another, and a few people get counted two or even three times; the net result of this is that the Census undercounts people.

To counter this, shortly after the Census, the ABS runs a post-enumeration survey - they sample people, match them to Census responses, and use this to estimate things like overcount and undercount (for instance, when the PES finds people who weren't recorded on Census), as well as various other quality-control purposes.

Census doesn't record people who were overseas at the time, but we're interested in them, so the ABS estimates the number of Australian nationals who are out of the country, and adds them towards the population.

The result of this is that a year or two after the Census was run, the ABS has a pretty good idea of the resident population on Census night (most recent being mid-06).

What if you want to know the population for June 30, 2007? To get a high-quality estimate ('estimated resident population', or ERP), you want to look at net births, deaths, and migration since then. Ideally you would just count registrations of births, death, and migration, and adjust by the net total - but in practice, not all of this data is recorded in a timely fashion, so a bit more analysis is required, involving various assumptions and corrections. (This can be quite tricky - for instance, because people needed to register births to qualify for the 'baby bonus', there was a spike in birth registrations at that time.)

This process takes time, so again you can only get these estimates in hindsight, and they're adjusted as more data comes in - the estimates for June 2007 won't be finalised until some time around 2012 or 2013, when the data from the 2011 Census has been processed.

If you want to estimate the population in the future, you have to start with a recent good estimate (e.g. the latest ERP), and make some assumptions from there about what births, deaths, and migration might look like in the future. Since the ABS staff aren't magicians, they can't do that with anything approaching 100% accuracy, so they produce a large range of possible values for these variables. (As the ABS disclaimer on this data notes, "The projections are not predictions or forecasts, but are simply illustrations of the growth and change in population which would occur if certain assumptions about future levels of fertility, mortality, internal migration and overseas migration were to prevail over the projection period.") Combining possible choices for those assumptions gives a large number of scenarios (currently 72); of these, the ABS picks three based on mid-range, high-range, and low-range assumptions. In the ABS' words, "These are intended to illustrate a range of possible future outcomes, although there can be no certainty that any particular outcome will be realised, or that future outcomes will necessarily fall within these ranges."

If you want to estimate the population right now... you have to do the same thing, because data on today's births/deaths/migration won't be available for quite some time yet. That's how the 'population clock' figures work, using the assumptions made for Series B.

The result of all this is that the 'population clock' figures - like any other projections - are not guaranteed to be particularly accurate. To be blunt, the population clock is there to give the casual public something they can look at without having to read through a lengthy description of what we really know (or don't) about the Australian population. If you want high-quality figures for the past, you need to use the ERP figures. If you want high-quality figures for the present, or the future, you need to wait until it becomes the past... and then go to the ERP figures. (And if you need to make decisions about the future now, you'd better be good at dealing with data where even the degree of uncertainty is uncertain.)

When you combine that uncertainty with the question of how exactly you choose the scope for a population, a discrepancy of around 550k is not particularly remarkable. Maybe the CIA picked a different set of assumptions for projection purposes. Or their update schedule means they're working off an earlier ERP figure. Or the 'population' they're counting excludes some category of people who are included in the ABS projections (e.g. Australians who don't live in Australia, or non-citizens who do.) It doesn't mean the Factbook staff are incompetent, just that they made some different choices when dealing with difficult issues requiring a lot of professional judgement.

What would be a mistake is to attempt to combine the two sources, using one as a guide to how the other should be interpreted, because there's no guarantee that their working definitions are consistent with one another's, and neither source was designed to be compatible with the other. --GenericBob (talk) 00:31, 17 April 2009 (UTC)

To be honest, I didn't read all of the above because there's just too much waffle there. I did skim it and it's obvious that you've missed the point, again. The ABS is the agency that collects all of the census data and issues the reports on the data collection. Those reports are used by every government and most non-government agencies, as well as international agencies. You won't find international agencies doing their own census. They use the data, and usually that means the ABS reports, for information about Australia. One of the tools that the ABS provides is the population clock which is reasonably accurate, despite it just being a projection. You can see this when new data from a census is released. The ABS explains how it arrives at the population data, the CIA doesn't. My point was that the CIA projection is so significantly (a lot more than 0.2%) different to the official figures that it lacks credibility. --AussieLegend (talk) 00:52, 17 April 2009 (UTC)

"the population clock which is reasonably accurate" - what do you believe the margin of error on the population clock to be, and what is your basis for that belief? When the advertised accuracy of a figure is 0.01% and the discrepancy is 0.2%, that's highly significant; when the discrepancy is 500k and the source makes it clear that they don't know the margin of error, it's quite impossible to judge whether the discrepancy is meaningful.

"Those reports are used by every government and most non-government agencies, as well as international agencies. You won't find international agencies doing their own census" - However, as I explained above, Census figures are only the first step in quite a long chain of work to produce current population projections. In this case, as noted in the Factbook documentation, the CIA's figures begin with census data - undoubtedly sourced by the ABS - but after that, the ABS' US counterpart (Bureau of the Census) does its own projections - probably making a few different assumptions along the way. I'm sorry if the explanation is a bit dry, but if you want to understand why two figures might disagree, you need to know how those figures are calculated.

You've also ducked the issue of scope differences. The grey areas in defining who we count as the 'Australian population' amount to far more than 500,000 people, so even if by some miracle both ABS and BotC had perfect and current data for the entire world population, you can expect see discrepancies of this size depending on exactly who each organisation decided to count towards their totals. --GenericBob (talk) 01:39, 17 April 2009 (UTC)

Okay. So your argument for using 'European' as an ethnic category for Australia, in data derived from an ABS source, requires going to a different source, disregarding the categorisation used for 'Australia', and instead seeking out the data for a small offshore territory - one with a unique history that makes different categorisations relevant - and then taking this classification scheme back to the ABS data for Australia. With no assurances that the CIA calculations used the same methodology (and in fact some pretty clear evidence that they didn't, in that discrepancy for 'Asian' figures). And the CIA Factbook is untrustworthy in how it categorises Australia's ethnic groups because "what it says is wrong" (based on your comparison of a different data item to a primary source on another site), but the page on Norfolk Island is trustworthy - not just as a representation of NI, a small island of about 2000 people far out to sea, but also when you extrapolate its approach to a country of twenty million people. Do you really not see this as WP:SYNTH?

Reading broadly is great. Digging through sources to find the slightest hooks to support your case - while ignoring more prominent and relevant bits in those same sources that oppose it - is not. If you're serious about putting properly-sourced data in this infobox without OR... well, why not just copy the "ethnic groups for Australia" data across directly from the CIA Factbook? We're agreed that it's a good quality secondary source, and it specifically refers to 'ethnic' groups - we don't have the 'ancestry' issue to deal with. That would be far better than rummaging through the Factbook to find a reference to a small sub-entity of Australia and then transplanting the categorisation scheme to another source's data - unless the point of the exercise is to defend your previous edits rather than considering whether they're actually appropriate. --GenericBob (talk) 21:02, 16 April 2009 (UTC)

For those who aren't familiar with Norfolk Island, it's an external territory of Australia - while nominally part of the country, it's mostly self-governing (which is presumably why it has its own separate entry in the Factbook). A lot of its inhabitants are descendants of the Bounty mutineers, and there's a significant social distinction between them and much more recent arrivals from Australia and NZ - which is probably why the Factbook has chosen to divide it into "descendants of the Bounty mutineers, Australian, New Zealander, Polynesian". For choosing categories for an article on Australia, it seems more reasonable to go to the entry for something more representative for Australia overall... like, say, 'Australia'. --GenericBob (talk) 21:12, 16 April 2009 (UTC)

@AussieLegend, 1. the CIA world factbook doesn't list "Australian" as an ethnic group. 2. Encyclopedia Britannica may interchange the terms, but wikipedia does not. Removed Again Intranetusa (talk) 01:42, 20 April 2009 (UTC)

1. The CIA does indeed list "Australian" as an ethnic group. It doesn't matter that it's not listed next to Australia. It's listed somewhere in the Ethnic Groups column which means it's an identified ethnic group.
2. That's a very poor argument. Wikipedia relies on sources like Encyclopaedia Britannica for its information. If such a credible source interchanges the terms then there's no reason why we should ignore it unless we can find a more credible source that says Brittanica is wrong.
3. I notice that somebody reverted you. There's a message there. --AussieLegend (talk) 07:16, 20 April 2009 (UTC)

Seeing as that someone reverted with the comment "What does the CIA have to do with this?", I think the message is that it was a bad reversion by someone who was not aware of the talk page discussion. Intranetusa should have referenced the talk page, but still the revert was ignorant of this discussion. AKAF (talk) 07:30, 20 April 2009 (UTC)

You've selectively quoted from the edit summary, which actually states "You are being premature. This is being discussed on the talk page. And what on earth has the CIA got to do with it?". Quite obviously the editor was well aware of the discussion, despite your assertion to the contrary. --AussieLegend (talk) 07:50, 20 April 2009 (UTC)

The fact that the editor was unaware of the only valid source so far proposed for the ethnicity data makes me think that they were not up to date. AKAF (talk) 07:56, 20 April 2009 (UTC)

That isn't a fact - you're just making assumptions. You have no idea what the editor was or was not aware of. --AussieLegend (talk) 08:32, 20 April 2009 (UTC)

"Australian" is not an ethnicity, but neither is "White" or "Asian" or "Aboriginal" --they're all groupings of multiple races. The problem with categorising demographics in Australia is that a serious amount of people don't know where their ancestors came from or come from a mix of ethnic backgrounds... Can the words "ethnic groups" not simply be changed to something else (or does it come with the template thing)?

As for the sources, why is the argument about which is more credible even coming up? The ABS has actually conducted the research--it's a primary source and an internal agency, whereas the CIA (which is a foreign organisation based in the United States) is a secondary source that gets its info from other sources. Unless of course the postman lost my copy of the CIA's "tick the box next to your ethnicity" pamphlet... So if both sources are just as incongruous with the subject, why would you not automatically go with the primary source? That is if you can't change the wording of "ethnic groups"...

Night w (talk) 08:13, 20 April 2009 (UTC)

Wikipedia prefers secondary and tertiary sources. See WP:PRIMARY. I would point out that this is a policy rather than a guideline or essay. It is not up for discussion. AKAF (talk) 08:27, 20 April 2009 (UTC)

The key word there is "prefers". "Primary sources that have been reliably published (for example, by a university press or mainstream newspaper) may be used in Wikipedia, but only with care, because it is easy to misuse them" is what the policy actually says. I know you're going to bring it up so I'll preempt you. The policy also says "Any interpretation of primary source material requires a reliable secondary source for that interpretation." There was no interpretation. It was a simple summary of the data in the source based on geographical origin of the groups listed. Summarisation was required because there was too much information to place in the infobox, as I've stated previously and had to re-iterate far too many times. --AussieLegend (talk) 08:42, 20 April 2009 (UTC)

I think there is more than sufficient discussion above to establish that you have neither the training nor inclination to analyse that data source with sufficient care. It is further clear that you misused that data source (WP:SYNTH) by performing an analysis which was only weakly supported by the primary data. In contrast, the analysis of the CIA world factbook was performed by trained demographers, and is a good secondary source. This is a clear case where WP:PRIMARY applies, because the primary source does not say what you want it to, and WP:SYNTH does not allow an analysis of the sort you performed. I think it is extremely clear to everyone from the long threads above that the analysis you want to perform is far from simple. Further it is extremely questionable whether it is correct and whether you have the training to understand why it is not correct. I say this with the greatest of respect for your energy and other contributions, but the simple fact that you indicate the percentages to 0.01% accuracy is a very clear indication that you do not understand the data analysis at all. Therefore I would strongly suggest that in fact this is a textbook case of why WP:PRIMARY exists in the first place. AKAF (talk) 09:23, 20 April 2009 (UTC)

"have neither the training nor inclination to analyse that data source" - Exactly what part of "no analysis" and "simple summarisation" do you not understand?

"In contrast, the analysis of the CIA world factbook was performed by trained demographers" - It was? Who analysed it? As far as I can see you just duplicated what was in the entry. Is this your definition of analysis?

"WP:SYNTH does not allow an analysis of the sort you performed." - Again, for the umpteenth time, there was no analysis, it was a simple summary. There was a single source so, again for the umpteenth time, WP:SYNTH does not apply. - "Do not put together information from multiple sources to reach a conclusion that is not stated explicitly by any of the sources." - Note "multiple sources".

"I think it is extremely clear to everyone from the long threads above that the analysis you want to perform is far from simple." - You think that simple addition is not simple?

"the simple fact that you indicate the percentages to 0.01% accuracy is a very clear indication that you do not understand the data analysis at all." - Absolute and utter rubbish. This is yet another bad assumption. The only reason I used that precision is because I fully expected somebody would take issue if the percentages didn't add up to 100. I never once thought that somebody would take issue with people from Europe being labelled as European.

"Therefore I would strongly suggest that in fact this is a textbook case of why WP:PRIMARY exists in the first place." - I strongly suggest that this is a misinterpretation of policy. --AussieLegend (talk) 10:10, 20 April 2009 (UTC)

I am sorry that you appear to be having some trouble understanding the arguments of everybody else here. However I cannot be responsible for reiterating the same arguments for the nth time if you refuse to educate yourself on how data analysis, citation and demography work. I would also suggest error propagation and significant figures as further topics of study, and your local library should be able to help you with a good text. AKAF (talk) 10:57, 20 April 2009 (UTC)

"There was a single source so, again for the umpteenth time, WP:SYNTH does not apply." You have invoked both Britannica and the CIA Factbook to justify your analysis of ABS data. That makes three sources.

Have a look at the example of synthesis on WP:SYNTH: "(1)Smith claimed that Jones did X. Jones replied that X is acceptable... (2)The Harvard manual does not call X 'plagiarism'." The policy page makes it clear that this is synthesis - even though both statements are sourced, assuming that (2) is relevant to (1) is OR. "To make the second paragraph consistent with this policy, a reliable source would be needed that specifically comments on the Smith and Jones dispute and makes the same point about the Harvard manual and plagiarism. In other words, that precise analysis must have been published by a reliable source in relation to the topic before it can be published in Wikipedia by a contributor." This is synthesis even though the editor has merely implied the conclusion ("Jones is not a plagiarist") by sticking two sourced statements together.

Smith accused Jones of X. The Harvard manual does not call X 'plagiarism'. This, per example above, is OR.

Smith described this data as 'ancestry'. The Harvard manual defines 'ancestry' as equivalent to 'ethnic group'. Still OR.

The ABS described this data as 'ancestry'. Britannica defines 'ancestry' as equivalent to 'ethnic group'. Still OR.

The ABS described this data as 'ancestry'. Britannica doesn't actually define 'ancestry' as equivalent to 'ethnic group', but it uses it that way. OR as above, plus a layer of interpretation (and not a very safe one, at that - if you see a doctor smoking, would you take that as medical advice to go out and light up?)

The ABS described this data as 'ancestry'. Britannica doesn't actually define 'ancestry' as equivalent to 'ethnic group', but it uses it that way, so the ABS' data is equivalent to 'ethnic groups'. OR+ as above, plus a little bit more OR because now we're explicitly stating a conclusion from that synthesis instead of just insinuating it.

This is ABS data on ethnic groups. - OR++ as above. Leaving out the bits where the OR happened doesn't make it non-SYNTH, it just makes it badly sourced SYNTH.

"You think that simple addition is not simple?" - I'm not sure how you get percentages using only addition, but leaving that aside temporarily, I'm willing to acknowledge that the techniques you're using are simple in a couple of senses - they're naive, and they're easy to implement. The problem is that assessing whether they're correct is an entirely different question, and not a simple one at all. (But FWIW, the answer is an emphatic 'no').

If I tell you that Scot has three kids and Erin has two, and ask how many kids Scot and Erin have altogether, a primary schoolkid can calculate 2+3=5 - it's a very simple calculation. But that doesn't mean it's right... --GenericBob (talk) 03:05, 21 April 2009 (UTC)

"You have invoked both Britannica and the CIA Factbook to justify your analysis of ABS data. That makes three sources." - Wrong! Back in December, when I origianally summarised the information, I used a single source. The data has been modified since then, as different documents appeared at the ABS link but there was always only one source used for each change. The CIA Factbook wasn't referred to until 5 days ago in response to Internetusa's claim that Australian wasn't an ethnicity. Neither that or Britannica was ever referred to in the past and they were only used on 16 April to show that other reliable sources referred to those ethnicities, demonstrating that the usage is widespread. Good try but no cigar.

"I'm not sure how you get percentages using only addition" - There's some irony in the fact that before I settled on "The only reason I used that precision is because I fully expected somebody would take issue if the percentages didn't add up to 100", I had written, "The only reason I used that precision is because I fully expected some pedant would take issue if the percentages didn't add up to 100", but I thought pedant was too stong a word. I see now that changing the word was a mistake. Next time I'll stick with my gut. Percentage buttons are standard on even the most basic calculators today. Calculating percentages is no harder than simple addition. I can show you if you want.

"If I tell you that Scot has three kids" (etc) - Nice example but it's considerably more complex than summarising the data in the source and somewhat disingenuous of you to try using it as an example of something completely different. Again you've missed out on that cigar. --AussieLegend (talk) 05:52, 21 April 2009 (UTC)

Now please think about this, because I've put a lot of effort into a finding a simple example which explains the problems with your analysis of the data: If I tell you that at the zoo yesterday there were 100 monkeys, who consumed 50 bananas, 60 oranges and 20 watermelons for a total of 130 pieces of fruit. Assume that each monkey ate no more than one of each type of fruit. Your analysis method tells me that 38.46% of monkeys ate bananas, 46.15% of monkeys ate oranges and 15.38% of monkeys ate watermelons. There are several problems with this analysis:

1. Pieces of fruit is only weakly correlated with number of monkeys (as responses are only weakly correlated with persons)
2. You assume that the sets are independent (One monkey may eat more than one piece of fruit/one person may give more than one response)
3. At best it will be an approximation to reality, with an additional systematic error correlated with the size difference between the two sets (The size difference is 23% in this case, 17.6% for the census data)
4. It is particularly bad when you start to group sets. For instance to state that 53.85% of monkeys ate no oranges is a simple addition of the type which you have performed, but this has absolutely no basis in reality. Since each monkey ate no more than one of each type of fruit, then there are only 40 monkeys (40%) who actually ate no oranges. This is an error of 25.7% (of 53.85%).
5. The stated accuracy of the computation (plusminus 0.01%) is just plain wrong. For this computation a realistic number would be plusminus one monkey (1%).

Now please think about this, because I've put a lot of effort into a simple example which explains why your analysis method is flawed. AKAF (talk) 06:35, 21 April 2009 (UTC)

I had a longer response, but the editor ate it, which is perhaps just as well... while we could continue this argument until the end of time, I don't want to do that just for the sake of arguing. Whatever the problems may have been with content, I believe you (AussieLegend) have provided a solution in the form of a reliable secondary source that lists "ethnic groups of Australia" in a format suitable for the infobox, with no need for translation. You've also found an error in the ABS documentation, making it possible to get that fixed. And you've honoured your agreement that the methodology would have to change if the documentation changed. While I don't think we're going to reach agreement on the details of this particular issue, I would prefer to focus on a productive working relationship rather than continuing to thrash out an argument that's no longer contributing to the article. If we're all willing to accept the CIA Factbook figures here, I would be very happy to let the rest of this drop. Any objections? --GenericBob (talk) 07:47, 21 April 2009 (UTC)

I'm not entirely happy with the CIA data because I think it's problematic. In fact I'd rather see nothing there now but I'm not going to push for it. Let's leave things as they are and see what happens. --AussieLegend (talk) 08:16, 21 April 2009 (UTC)

If I could just return to the response to my talking about primary and secondary sources, if it's Wiki policy to go by secondary sources, why do the "ethnic groups" listed on the pages for Canada, New Zealand, the United States and the United Kingdom all go by statistics provided by their own government census bureaus? Is Australia so incompetent at gathering our own statistics that we have to rely on the "intelligence" of the CIA? Is that what you're saying? I don't like the CIA figures because I find them vague and offensive. I'm with AL on either changing them or removing them from the box entirely. Night w (talk) 08:20, 22 April 2009 (UTC)

It's not a question of competence. The ABS does not provide the statistics for the ethnicities in the form they are presented in the box (or noone has been able to find them). The data provided by the ABS requires further analysis. It is not a question of incompetence, but of preparation of the raw data. The analysis of demographic data is hard to do and requires trained people to do it, and the only source we could find which has done this analysis in the form which is useful for the infobox is the CIA world factbook. Aussielegend performed an analysis on the data which was provided by the ABS to get those older figures. This analysis has been shown by many people to be fatally flawed in a number of key areas.

Now, your dislike of the CIA aside, it does have competent analysts. I'm sorry that you prefer the outcome of Aussielegend's analysis, but his results were bollocks. I cannot speak for the other articles, but probably the relevant bureaus did the statistical analysis in the form required. If not then we prefer to use a good analysis, and saying that they used a poor analysis is no excuse for doing do here. Relevant pages are WP:OTHERSTUFF as referenced in WP:WAX.

If this really bothers you, why don't you write to the ABS and ask them to provide a good set of data for the Wikipedia infobox? This would actually be an optimal outcome, and maybe they can be persuaded. The data in that infobox is also very questionable in other areas, for instance the population is probably also not correctly cited (it is uncorrected population data), and the GDP comes from the IMF. AKAF (talk) 08:59, 22 April 2009 (UTC)

I haven't checked the others, but Stats NZ publishes ethnicity data (see e.g. [50]), based on a Census question that specifically asks "which ethnic group do you belong to?" (Note that their published percentages add to more than 100%, because they allow more than one response.) The Australian Census doesn't have an equivalent question; while I can think of at least six that are relevant to ethnicity (Indigenous identification, COB of person, COB of father, COB of mother, ancestry, and language), you'd have to put them together and apply some expert judgement in order to form them into ethnic categories. And as AKAF notes, you'd probably want to correct for nonresponse, since some groups have higher rates of nonresponse than others.

The fact that SNZ collects that information and ABS doesn't isn't about competence, it's about finite resources and priorities. The more data you collect, the more it costs to process; this caused major embarrassment in the 1976 Census, when the ABS got too ambitious and half the forms had to be shredded unprocessed. Even if you have an unlimited budget for processing, asking too many questions gets you to a point of diminishing returns (people get bored and don't finish the form). So everybody has to make hard choices about what to leave out. The NZ design is better suited to describing the country's ethnic groups, but it's weaker on other scores - for instance, it's easy to identify second-generation Australians, but AFAIK there's no way to pick second-generation New Zealanders. --GenericBob (talk) 12:40, 22 April 2009 (UTC)

Developed country?

The statement: "Technologically advanced and industrialised, Australia is a prosperous multicultural country and has good results in many international comparisons of national performance such as health care, life expectancy, quality-of-life, human development, public education, economic freedom, and the protection of civil liberties and political rights." is quite misleading.

First - Australia has very poor internet network comparing to many countries in the world. Secondly, a country with slow and in some instances still not electrified trians cannot be technologically advanced. Industraliazed also is misleading - a country that is predominantly a primary producer, that doesn't manufacture even one bicycle, TV etc - cannot be called industrialized.

Multicultural? - just because there are many migrants - it doesn;t mean it is multicultural - when de facto it is an English colony which enforces English language at schools and English way of life.

Freedom with internet censorship cannot coexist. Democracy while still monarchy? Do you remember Whitlam Government??? Is this a democracy when the monarch can dissolve a government within one day????

Health care with long queues to emergecy units, and long queues for basic public health services such as dentist (usually people wait 2-3 years) is not at the top of the world.

Life expectancy - especially is local Aboriginal communities is very low.

Human development - still no bill of rights.

Public education - very low, no foreign languages at school, ranking very low in maths among other countries.

Hope you can amend this sentence and be true about Australia —Preceding unsigned comment added by 115.129.17.160 (talk) 01:45, 19 April 2009 (UTC)

Woah--where do you live??? Australia has a higher percentage of Internet users per capita than Japan, the UK, the US, France... AWA Limited manufactures televisions and, I don't know about bicycles, but haven you ever heard of Holden?

The government doesn't enforce English culture--freedom of religion and society is guaranteed in the constitution and you only have to walk down a city street to notice multiculturalism. I live down the road from a mosque. And just because the GovGen can exercise dismissal powers (that have only ever been used in an emergency), doesn't mean she can impose direct rule--the people choose the government. On the Democracy Index, Australia is ranked among the top 10 most democratic countries in the world (alongside seven other monarchies) by The Economist.

I don't know where you line up for your health care provider, but is it not free? It's very expensive in most countries. Check this research, which ranks Australia's health care system above that of both Canada and the US. As for average life expectancy, Australia is ranked 5th by the UN. So it's not low "especially" in Aboriginal communities, just specifically in those communities.

Did you go to the "Developed country" page before you Wikilinked it for this section's title? Australia is ranked 4th...

We don't live in the Dark Ages down here...but I think you knew that, and were just ranting. If you're not happy with the level of freedom and development in Australia, then move to Scandinavia. Oh...but they're all monarchies....

Night w (talk) 07:29, 20 April 2009 (UTC)

Anon commenter, it might help to consider issues of geography. Australia has about the same population as Sweden, Denmark, and Norway combined, spread across roughly ten times the land area. This has a lot of influence on what's practical for infrastructure projects (e.g. rail and communications). --GenericBob (talk) 23:49, 21 April 2009 (UTC)

[removed remainder of thread, which had ceased to be even tenuously related to how to improve the article. Hesperian 04:42, 24 April 2009 (UTC)]

ref improve or remove

The following three lines are from the article:

• Australia willingly participated in World War I. [Ref: Bean, C. Ed. (1941). Volume I - The Story of Anzac: the first phase, First World War Official Histories, Eleventh Edition.]

Not supported by linked page, is Bean 1941 the reference? Conscription suggests that the inhabitants of 'Australia' were less enthusiastic!

• Many Australians regard the defeat of the Australian and New Zealand Army Corps (ANZACs) at Gallipoli as the birth of the nation—its first major military action. ["Rudd angered by Gallipoli remarks". BBC News. 1 November 2008. Retrieved 2 November 2008.]

This swiped text is from a BBC article that contrasts the views of a former Prime Minister with the current one. Should we add ",... and many don't"?

• The Kokoda Track Campaign is regarded by many as an analogous nation-defining event during World War II.

Uncited! Analogous?

– cygnis insignis 05:33, 19 April 2009 (UTC)

I added the appropriate cites from the recent scholarship, especially for Kakoda see Hank Nelson, 'Gallipoli, Kokoda and the Making of National Identity', J of Australian Studies online Rjensen (talk) 14:03, 19 April 2009 (UTC)

More likely, WP:FAR YellowMonkey (click here to vote for world cycling's #1 model!) 05:35, 19 April 2009 (UTC)

I agree with the 3 dot points. I also suspect that FAR may be in order. The risk is that once the FA status is questioned or removed, attempts to improve it could go the other wasy - ie, all bets are off and nothing is sacred. As major changes to the culture section over the last few months suggested, I suspect the article could lapse into a state of even less quality. At least the current FA status is a way to hold off decrease in article quality. Hmmmm --Merbabu (talk) 06:35, 19 April 2009 (UTC)

I disagree will all three dot points.

• Supporting participation in the war and supporting conscription are two entirely separate issues. It is possible, and indeed highly likely than many people who opposed conscription (including servicing soldiers who voted) still supported Australia's participation in the war, with volunteer enlistment. This is neither contradictory or unusual. The war effort was popular and bipartisan and tested through elections. Casting Australia as an unwilling participant is historical revisionism. -- Mattinbgn\^talk 01:27, 20 April 2009 (UTC)
• No, the fact that "many don't" is obvious from the "Many Australians..." rather than "Most Australians..." or All Australians ..." This is nit-picking.
• The Kokoda campaign was put forward by Keating as an new and better "Gallipoli" Sources for this should be easy to find. -- Mattinbgn\^talk 01:27, 20 April 2009 (UTC)

Let's not go too FAR :) There are many fine editors active there, but this is a the top priority article for a wikiproject with some of the best. cygnis insignis 13:09, 19 April 2009 (UTC)

Yeah, if preventative action is taken it will be all fine and dandy, like when I pre-emptively referenced flag of Australia last year. There is a problem that if people make a lot of ad hoc edits the prose can become disjointed likke a list of trivia though. YellowMonkey (click here to vote for world cycling's #1 model!) 01:13, 20 April 2009 (UTC)

anon edits

it seems every single one needs to be reverted because of vandalism. can we get this page locked again? Rehumanist - talk 12:58, 26 April 2009 (UTC)