Talk:Campus sexual assault

RfC on recent AAU campus climate survey

The following discussion is an archived record of a request for comment. Please do not modify it. No further edits should be made to this discussion. A summary of the conclusions reached follows.

No consensus. It's a shame to close an RfC like this, but despite a lot of time and noticeable effort at getting participation, there have been only 3 users that actually support either option. --GRuban (talk) 20:20, 28 December 2015 (UTC)

---

There is a debate on how we should include a widely reported statistic relating to the percentage of victims, who when asked why they didn't report the incident, answered they didn't think it was "serious enough" per the study and news reports.

The section leads with a high level statistic for the rates of all types of sexual assault and misconduct, "The 2015 Association of American Universities (AAU) Campus Survey on Sexual Assault, one of the largest studies ever of college sexual violence, drew responses from 150,000 students at 27 schools, including most of the Ivy League. It found that more than 20 percent of female and 5 percent of male undergraduates said that they were victims of sexual assault and misconduct".

The AAU study, Table 6-1, page 110 contains the following:

Percentage reporting or not, with reason	Penetration by Force	Penetration by Incapacitation	Sexual Touching by Force	Sexual Touching by Incapacitation
Contacted at least one program in university list	25.5%	13.3%	7.0%	5.0%
Did not contact any programs reason: "I did not think it was serious enough to report"	58.6%	62.1%	74.1%	75.6%

Several news reports that covered the high level findings in AAU survey also included the following details from page 110. I have provided a few samples (with direct quotes provided):

• Washington Post, "the vast majorities of poll respondents who said they had not reported to campus authorities the events that the AAU classified as sexual assaults — 'the dominant reason was it was not considered serious enough,'
• Chicago Tribune, "More than half of those who didn't report a rape said they didn't think it was serious enough to report."
• NJ.com "The most common reason cited by students for not reporting an incident was that they didn't consider it serious enough"
• CNN "More than 50 percent of the women who reported some of the most serious incidents, including forced penetration, didn’t report it because they didn’t think it was “serious enough,” according to the survey."
• The Daily Toreodor "More than 50 percent of the victims of even the most serious incidents, such as forced penetration, said they did not report the event because they do not consider it “serious enough,” according to the survey."
• Campus Safety Magazine, "When students were asked why they did not report incidents of sexual assault and sexual misconduct, the most common reason was that it was not considered serious enough."
• My Fox 8 "More than 50 percent of the women who reported some of the most serious incidents, including forced penetration, didn’t report it because they didn’t think it was “serious enough,” "

Looking for input on two approaches discussed in this talk page to handle this:

1. In the 2015 Campus Climate Survey section, add language with some of the multiple cites including the AAU study along the lines of, "The survey also reported that the majority of students whose responses were classified as sexual assault did not think their experience was "serious enough to report"."
2. Only include this finding in the 2015 Campus Climate Survey if it includes criticisms/counterpoints to these kinds of finding about surveys in general, making points like, "First, a salient research issue is what students mean when they define incidents as not serious enough to report. For conservatives, the phrase "not serious" is taken in a strictly literal sense as meaning that the incidents were unimportant. For feminists, however, such a response may merely indicate a false consciousness expressed by women acculturated to see their victimization as somehow acceptable." (a partial quote from an academic source provided by Nblund).Mattnad (talk) 13:56, 15 November 2015 (UTC)

• CommentMattnad, this seems to mis-characterize my perspective. I did not say we should leave this statistic out. What I said was that we should mention it alongside the previous research on this topic. Importantly, we should explain how social scientists generally interpret that result, and explain that researchers doubt that this finding indicates that these are not serious sexual attacks. I would appreciate it if you would edit this down for brevity (maybe put the citations and other commentary in the section below) and clarify the wording of option #2 to more accurately reflect what I proposed. Nblund (talk) 00:42, 8 October 2015 (UTC)

I think editors can sort it out with the citations since you several times argued that it was WP:Undue because there was only one source and you insisted that the Survey authors didn't highlight this, so it was not relevant to use it either. I think it's important to demonstrate how common this detail is given your past statements so other editors can see what's been going on here. I don't think option 2 misrepresent your position at all.Mattnad (talk) 01:35, 8 October 2015 (UTC)

I'm sure it wasn't intentional, but yes, it does misrepresent my position. I don't think we need to exclude this detail. I do think we need to contextualize the detail along with past research. I think it would waste people's time to request an RfC in which an option is offered that neither one of us supports. Change it please. Nblund (talk) 01:58, 8 October 2015 (UTC)

Change has been been made.Mattnad (talk) 11:38, 8 October 2015 (UTC)

RfC comments

• Approach #1 is neutral, reflects both the study and multiple reliable sources that covered this particular survey. Approach #2 comes from academic sources that do not specifically discuss this survey, and are grounded in POV that's WP:Fringe. Not a single news source I could find on the AAU survey addressed this counterpoint and to create a separate discussion would be WP:Undue. I will add that Nblund's reasons for objecting to this detail as part of the AAU study is that "I'm concerned that it makes the findings appear less impactful by misleading people". This concern was not considered material by professional news services covering the survey, or a plain reading of the survey itself. My take is Nblund would like to avoid reducing the perceived impact of the numbers by discounting or eliminating what survey respondents themselves stated.Mattnad (talk) 14:15, 5 October 2015 (UTC)
• Approach #1 I agree with Mattnad in his thorough analysis of the different approaches. Immortal Horrors or Everlasting Splendors 21:56, 5 October 2015 (UTC)
• Approach #2 provides greater context by explaining different interpretations of the finding. It does so in an even-handed manner, presenting the perspectives of both sides. I am mystified by the suggestion that it includes a fringe POV - how can a substantial body of academic sources (which are recognised as top-quality sources by policy) can be regarded as fringe? Neljack (talk) 21:55, 24 December 2015 (UTC)

Comment: For context, its important to note that this particular finding is not unique to this survey. This exact question has been asked on previous studies of sexual assaults, and the result is consistent with previous findings. Researchers generally attribute the response to minimization on the part of victims, and note that victims of even very serious attacks sometimes decide it isn't worthwhile to report the attack to the police. (see: Fisher, Daigle, Cullen & Turner 2003) Some critics of these studies (such as Christina Hoff Sommers) have suggested that the "not serious enough to report" response indicates that the concerns about sexual assault on college campuses is overblown, but that view is not supported by the empirical evidence, and experts like Bonnie Fisher are generally skeptical of this interpretation (see the quote below)

B.S. Fisher:

Recall the study of college students who found that robbery victims also did not report their victimization because it was “not serious enough” (Fisher & Cullen, 1999). It is doubtful that critics would conclude that robbery victims were a methodological artifact created by politically correct researchers. Second, we need to probe more deeply into what seriousness means to victims. It appears that it involves suffering an injury and the presence of a weapon. Acquaintance rapes do not typically involve these features. Third, the female students did not say that their victimization was not serious but not serious enough. This standard suggests that reporting a sexual victimization to the police involves a cost—a loss of privacy, potential embarrassment, having to “deal with” one’s parents, rejection by friends of the perpetrator one accuses, the necessity to perhaps leave the campus and drop out of college, and having to testify at a college disciplinary hearing or court case. In this context, the harm experienced—especially in the absence of visible physical injury—may not seem serious enough to pursue an assailant legally.

Again, I'm not saying we should exclude this finding, I'm saying we should cite it alongside the previous findings and alongside previous research on the topic. If this particular detail is important enough to mention, why isn't it important enough to mention the views of experts on the topic? Wikipedia is not a newspaper, or a random collection of events. The goal here is to inform people, and the notion that we should suppress obviously relevant, reliably sourced, information strikes me as fundamentally anathema to that project. Nblund (talk) 00:42, 8 October 2015 (UTC)

While critics may point to this, Approach 1 makes no statement about the whether or not the survey was overblown. And CNN, NJ.com, Campus Safety that include it do not either. It's a neutral presentation ideally using the survey itself as the source. I think your concerns are political rather than encyclopedic.Mattnad (talk) 02:21, 8 October 2015 (UTC)

• Comment: What a shame this study's question was worded so poorly. "Not serious" could have been taken by respondents as meaning "not something that bothers me" or as "not something anyone else would take seriously if I reported it" (or more narrowly as "not something authorities would take seriously if I reported it") or any number of other interpretations. And it's impossible for anyone to know what the women who took part in the study thought that vague phrase meant. If previous studies used the same poor wording, that wasn't sufficient reason to repeat their mistakes... but that's outside our purview here. Coverage of the study in reliable sources should of course be included, but effort should perhaps be taken to find an examination of the study's methodology in a reliable source or sources so that this issue can be included in the text of the article as well. —GrammarFascist _contribs^talk 18:07, 27 October 2015 (UTC)

Welcome to the world of these kinds of studies. This isn't the only example of ambiguous questions. In regard to this particular study, there are direct criticisms of the questions used to capture sexual assault as well. In the interest of parsimony, we haven't included those either.Mattnad (talk) 20:17, 27 October 2015 (UTC)

• Comment: 1. The long list of news articles demonstrates the notability of the study, but Wikipedia shouldn't be summarizing the news article, but the published researchers' published findings. ("For information about academic topics, scholarly sources and high-quality non-scholarly sources are generally better than news reports." WP:NEWSORG) The researchers' summary is on page iv:

A relatively small percentage (e.g., 28% or less) of even the most serious incidents are reported to an organization or agency (e.g., Title IX office; law enforcement); More than 50 percent of the victims of even the most serious incidents (e.g., forced penetration) say they do not report the event because they do not consider it “serious enough.”

2. As a side point, the percentages reported here are fractions of those who "Did not contact any programs." They do not represent fractions of the number reporting assaults. (see p. 112)

3. The researchers publishing this report observe, "When asked why the student did not report an incident, the dominant reason was it was not considered serious enough. This is also consistent with prior research (e.g., Fisher et al., 2003)." (p. 36) So I think that Fisher 2003's explanation for this reporting (as quoted above) should be summarized and cited first. It can probably stand on its own, but following up with point-counterpoint a la version 2 isn't unreasonable either.

4. In my opinion, provinding greater detail about the other survey findings (as highlighted on page iv) is more encyclopedic, and a higher priority than a deep dive on this unclear "not serious enough" qualifier.--Carwil (talk) 01:48, 11 November 2015 (UTC)

• Comment I re-posted the RfC tag for this discussion, since there doesn't seem to be a clear consensus still. Carwil correctly notes that the AAU authors note that these findings are consistent with Fisher et. al 2003 -- which is the source of the quote mentioned in the original RfC post above. I think this should resolve Mattnad's stated concern that adding this material would constitute original research. Given that the authors directly contextualize their findings in light of previous work, it seems fairly clear-cut that the additional context and critique from Fisher ought to be included. Nblund (talk) 01:06, 17 November 2015 (UTC)

Fisher (1999) is not the same as Fisher (2003), p. 36. So whatever they are referring to in the study (Fisher 2003), it's not the quote provided by NBlund (Fisher 2009). Perhaps there's a linkage, but two different books are not the same book.Mattnad (talk) 02:48, 17 November 2015 (UTC)

I've cited two different Fisher quotes. The quote you paraphrase at the top of this RfC is from Fisher et. al 2003. That 2003 paper is the same one cited by the authors of the AAU study. The other quote comes from Fisher's 2009 book, Unsafe in the Ivory Tower. That book is not cited in this paper, but it's a discussion of the same topic by the same (highly respected) author cited in the AAU report. It's really tough for me to see how it would be undue to cite the research that the authors themselves discuss. Nblund (talk) 03:24, 17 November 2015 (UTC)

• Comment I feel a little silly for not noticing this earlier, but I just saw this: Bonnie Fisher is actually on of the lead authors of the AAU report (left column, second row). So it really seems like her views are pretty clearly relevant here. Nblund (talk) 15:49, 28 November 2015 (UTC)

I don't see how that on its own justifies bringing in her other, and very separate, opinions on the topic.Mattnad (talk) 15:41, 1 December 2015 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

RFC regarding AAU subsection.

There's an ongoing dispute over how to deal with an editorial argument in the subsection on the recent AAU report. Specifically, the section presents a statistic that most people who experienced a sexual assault who did not report the incident said that they did so because they "did not believe it was serious enough to report". It then presents this argument from an editorial by Stuart Taylor jr.: "Stuart Taylor, writing for the Washington Post, remarked "This most plausible explanation is that most of those classified by the survey as “victims” of sexual assault or rape did not really think that they had been sexually assaulted."" The question is over where/whether to include Taylor's comment, and over whether or not it is appropriate to note, in that section, statements from past research that challenge arguments similar to the one made by Taylor regarding the AAU study. A few solutions that have been proposed are:

• 1. Leave the entry as is, or use a similar argument from a different editorial such as this one by Emily Yoffe.

• 2. Move the quote and statistic to a separate subsection that addresses the causes of non-reporting, while also discussing past research on non-reporting, and the "not serious enough to report" response, and the criticisms voiced by people like Taylor.

• 3. Leave Taylor's argument where it is, but also cite past research that disputes Taylor's argument within the AAU section.

Please use the section below for comments. Nblund (talk) 18:43, 12 January 2016 (UTC)

Nblund did not include what I had proposed several times, that the AAU section has what reliable sources say about the responses, and she can create an seperate of what some others say (per Nblund's sources) about these surveys in general. As written, this RFC forces an either/or dichotomy, which was is not a compromise or reflective of what multiple sources say in 2015, rather than opinions from 2003. This would be a 4th option.Mattnad (talk) 14:30, 18 January 2016 (UTC)

I'm not sure I follow what you're saying closely enough that I feel comfortable summarizing it above, but you're welcome to add it as a fourth option in the portion above if you just sign it. Nblund (talk) 23:31, 19 January 2016 (UTC)

Comments

Option 2 Taylor is a columnist with no relevant research background presenting a viewpoint that has been explicitly rejected by experts. It seems to violate NPOV to grant article space to a minority opinion while leaving out the views of scholars. The argument is probably notable enough to warrant some mention somewhere, but it should also be made clear to readers that it's a view generally held by non-expert critics of these studies, and that experts view these results very differently from Taylor. Since Taylor's argument is one that has been made with respect to other studies discussed in the subsection (including the 2007 Campus Sexual Assault study and the National College Women's Sexual Violence Study), it's probably better to place the discussion in it's own subsection. Nblund (talk) 19:13, 12 January 2016 (UTC)

Option 2 (though Option 3 would also be reasonable). I agree with Nblund's comment. As I said above, this approach "provides greater context by explaining different interpretations of the finding. It does so in an even-handed manner, presenting the perspectives of both sides. I am mystified by the suggestion that it includes a fringe POV - how can a substantial body of academic sources (which are recognised as top-quality sources by policy) can be regarded as fringe?"[1] Neljack (talk) 19:54, 12 January 2016 (UTC)

Comment FYI: because of the low participation, I went ahead and posted a notice about this RfC on the NPOV and OR noticeboards, if needed, we might also post in the Wikiprojects for sociology or criminology/law. Nblund (talk) 23:49, 15 January 2016 (UTC)

Option 1 or a 4th option. There are ample secondary sources referring to the AAU study in 2015: [2],[3], [4],[5],[6],[7], [8], [9], [10]. The survey was also different from any past efforts in that included questions around affirmative consent. There are no secondary sources that refer to this. Rather, there's a 2003 opinion that Nblund wants to apply to this survey. At most, we can discuss the opinions on this matter separately given the lack of specificity to the AAU survey. I had several times proposed we keep the AAU section as is, and Nblund can create a separate section on these topics in general. This is a pretty basic WP:RS issue. I will add that one of the sources Nblund has cited dwells on Feminist opinions about the matter, and presents their views as hypotheticals. Per Fisher, "For feminists, however, such a response may merely indicate a false consciousness expressed by women acculturated to see their victimization as somehow acceptable." Fisher is not referring mainstream research here, but Feminist opinions. Finally, Fisher herself sees the survey as breaking new ground. Per a Washington Post Article on this, "The dominant reason for why students who didn’t tell authorities: They said it wasn’t serious enough. “That will stimulate a lot of discussion,” said Bonnie Fisher, a professor at the University of Cincinnati and a Westat consultant. “We as researchers don’t know a lot about this — it hasn’t been measured in the past.” [11] Nblund has argued that Fisher, as originally quoted didn't mean what she said, but no matter how you slice it, the AAU study is different from past according to Nblund's preferred expert.Mattnad (talk) 14:30, 18 January 2016 (UTC)

Option 1 or Option 3. You haven't linked the source in question, but I assume it's this? Taylor's qualifications are listed in that source, "Stuart Taylor Jr., an author, journalist and Brookings Institution nonresident senior fellow, is writing a book with KC Johnson about the alarm over campus sexual assault." I don't see in what way this would possibly make him unqualified to discuss campus sexual assault. This is a controversial subject and there are going to be opinions on both sides, as there have been throughout the article. Using a lack of opposing viewpoints as a reason to justify not including an opposing viewpoint would be manufacturing consensus by way of censorship. Taylor's views are in no way fringe or unqualified, so as long as they are presented and attributed as his opinion, they are as subject to additional support or criticism as any other statement in the article. Scoundr3l (talk) 20:42, 18 January 2016 (UTC)

Scoundr3l, option 2 suggests moving Taylor's critique. None of the proposed options include removing Taylor's criticism, or treating them as fringe views. The question is whether we can also cite the views of people who research sexual assault who have previously disputed Taylor's argument. Nblund (talk) 21:36, 19 January 2016 (UTC)

I apologize if I was unclear in my position or misunderstood the intention of option 2. My response was partially meant to address the concern that this is a non-expert opinion when I see no reason to treat it as such, based on the qualifications listed. To elaborate, based on the fact that it appears to be qualified commentary, I am generally opposed to moving the comment to a new subsection unless this subsection is for all second-party commentary on the survey data. The statement itself doesn't appear to be a direct interpretation of the survey results, so it doesn't necessary belong with other direct interpretations. However, given the qualifications, the brevity of the statement, and that fact that it seems to relate only to this survey (as opposed to such surveys in general) I believe a new paragraph would be sufficient in this case. Absolutely it should include relevant dispute to Taylor's views. If we're talking a substantial discussion, I could support option 2, but the majority of this article manages to discuss opposing views (even flip-flopping oppositions to the opposition) within the body of the particular section and I would think this comment could do the same. Scoundr3l (talk) 05:17, 20 January 2016 (UTC)

We could do a "criticism" subsection, but my idea was to make a subsection specifically about reasons for not-reporting sexual assaults. That section would cite academic research on non-reporting, and then cite Taylor's critique alongside other authors who have made this criticism regarding previous studies. Taylor is specifically discussing the AAU, but the finding he is citing is consistent with past research, and the argument he is making is an old one that has been applied to other research discussed in that section of the entry. For example, Cathy Young makes this same point about the 2000 NCWSV study here, while Asche Schow makes this criticism about the Michigan study cited in the AAU section here. Citing it only in the AAU section can give the misleading impression that the AAU is unique in this regard, when it really isn't. We could cite those arguments, and cite the response from academics. Nblund (talk) 18:28, 20 January 2016 (UTC)

Gotcha. I'm not sure I'll know exactly how strongly I can support option 2 until I see it, but I'm not opposed. It does appear to me that Taylor is not only providing commentary on reasons for non-reporting, but also criticism of these survey's results. For the former reasons, and because there are plenty of other such comments, I could conditionally support option 2 and see the value of that subsection. However, in the context of the quote as a criticism of the survey, I also feel that it belongs alongside other commentary related to the AAU survey, so I support including it there. Were the decision solely up to me, I would consider another option of including Taylor's commentary in both areas, though to avoid redundancy, I would suggest finding a different suitable quote from Taylor's article as it pertains to commentary on the AAU results. That's up to the editor. I hope that helps. Scoundr3l (talk) 21:59, 20 January 2016 (UTC)

Comment. I see some serious problems with this section, which likely render this RfC irrelevant, and I recommend it be closed.

My first instinct on reading this was to say that, since Taylor's is an opinion piece, it probably doesn't deserve a place in criticizing a scientific study. However, on careful reading I found that most of the section is based on even lower-quality commentary from newspapers. Per WP:NEWSORG, caution should be exercised when using such sources, and I don't think that has been done here. Consider the following claim:

The AAU’s findings are roughly consistent with a Washington Post-Kaiser Family Foundation national poll, published in June 2015, that found that 1 in 5 young women who attended a residential college during a four-year span said they were sexually assaulted.

This is based on newspaper reporting making this connection. However, the AAU report itself explicitly cautions against making this claim. (Page v of the executive summary.) This is an example of shoddy reporting. I think you ought to take anything written in newspapers by non-experts with a grain of salt.

Another thing which raised my eyebrow was the way the primary source was cited. Particularly this claim:

The AAU surveys also found that, 3.2% of undergraduates were victims unwanted anal, vaginal or oral sex that occurred because they were forced, threatened with violence or incapacitated and unable to consent (commonly known as rape) in the past year.

Because this was cited without a page number, I had to search through the document for over an hour in order to figure out where this came from. My best guess is that it is derived from Table 3-21 on page 82. However, the data there don't quite support the claim above, which has multiple inaccuracies and WP:OR issues.

(1) "Unwanted". The actual wording, as you can see at the top of page A5-23, is "nonconsensual or unwanted". (Moreover, in an issue Taylor might criticize, neither term appeared in the actual questions asked.)

(2) "Anal, vaginal or oral sex". This appears to be original research. The actual definition includes e.g. penetration with fingers or an object, and does not make reference to these specific terms.

(3) "Commonly known as rape". This is the most egregious of the problems, and also the most subtle: the report never says this. This notion likely derives from the description on page 11, section 2, which asserts that violations made possible through physical force or incapacitation as operationally defined in the study generally meet the legal definition of rape. That's not at all the same as being synonymous with the legal definition of rape, and there is no claim to the effect that all of the "yes" responses represent a rape as legally defined. (As Emily Yoffe's analysis points out, there couldn't possibly be.) Moreover, this datum in the table also includes attempted but not completed forcible penetration, which I think is legally "attempted rape."

The final thing which I found very strange about this section was the way Stuart Taylor's criticism was cited. He is referenced only for a relatively mild criticism of the report's conclusions, when in fact both he and Emily Yoffe are essentially saying that the whole report is straight-up pseudoscientific bunk. If you are going to cite these people's opinions, you might as well faithfully represent the main point of what they actually said. In Taylor's words the study "was itself deliberately designed to exaggerate the number of sexual assaults on campus" and is "grossly misleading" in multiple respects. He points out multiple serious methodological flaws and notes that, if the number of persons who said that they reported a rape were extrapolated, the result would overestimate the total number of actual sexual assault reports (not just rape) by a factor of nine. Emily Yoffe echoes this and other criticisms of his. These are much stronger statements than the one which this RfC concerns.

Taylor is writing a book on the subject and it might be best to wait for his book to be published and see how it is received. For now, I can see two ways to go.

A. Dramatically cut down the section, removing claims cited to newspapers, and be very careful to ensure that citations to the study conform to WP:PRIMARY by refraining from any original interpretation of the study. Do not present this non-peer-reviewed study as though it had equal weight with the much better-established NCVS.

B. Present the whole controversy, with full discussion of the harsh criticisms of this study's validity, and place it in context of the political turmoil surrounding the campus rape issue. This means restructuring the article in a way that places primary emphasis on, for example, the Dear Colleague Letter, the campus anti-rape social movement, and the cultural significance of the "1 in 5" statistic (regardless of its validity).

Obviously B would make the better and more informative article, if done right, but it would depend on editors with deep ideological differences trusting each other enough to ensure that this doesn't remain the WP:BATTLEGROUND that it seems to currently be. I have no opinion as to which of these is a better course of action, and I personally don't intend to be involved in this article. --Sammy1339 (talk) 00:24, 19 January 2016 (UTC)

TL;DR: Because of the fourth and fifth bullet points in WP:SCHOLARSHIP and the second bullet point in WP:NEWSORG, most of the contents of this section should be cut. --Sammy1339 (talk) 03:32, 19 January 2016 (UTC)

What you've noted are changes made by Nblund that have distorted the text so it no longer resembles the sources. For instance, the AAU survey explicitly does not use the term "rape" or "sexual assault" in its questions: per this Slate article, "The report deliberately does not use the word rape, Cantor told me. This was at the universities’ request, because the schools are addressing conduct violations, not criminal matters." [12]. So here Nblund inserts "rape" despite the term not being in any of the sources. When I objected, on the grounds it was not in the sources, Nblund just rephrased it. It's pure POV pushing frankly. Nblund wants to equate the criminal definition with broader and far looser definition relating to conduct even though the definitions are different, as are the terms used by the sources.Mattnad (talk) 14:00, 19 January 2016 (UTC)

For context, in case anyone else wants to follow this discussion, the claims I specifically discussed above were subsequently removed in this diff. --Sammy1339 (talk) 18:19, 19 January 2016 (UTC)

Comment Sorry, forgot to add the edit summary on those removals. Sammy1339, I think you raised a good point, and I went ahead and removed the portions tagged as dubious. I agree that the best option would be a rewrite of the section that avoided deep dive in to specific studies in favor of a substantially reduced version that laid out the relevant debates regarding measurement in brief and that placed the views of non-expert skeptics in a single subsection. I have an incomplete draft version on my sandbox, if anyone thinks this is a good alternative option, I'm down for it.

Yoffe's claim about the over-estimation of assaults is rooted in a misunderstanding of how Clery Act reporting works. I don't think that it really matters which specific argument from these op-eds that we use, because they all have the same basic problem. Yoffe and Taylor are generally skeptical of this entire line of research, not just the AAU, and I think it's a little silly to pretend that these completely untrained columnists are just motivated by a concern about sampling biases or survey methodology on any one particular study. Nblund (talk) 18:28, 19 January 2016 (UTC) Edit: I think I should note: the question about measuring rape and sexual assault is actually a scientific question, so making a major shift toward discussing the cultural importance of the stat regardless of it's validity seems a little like it strikes a false balance. It's really a debate between experts who generally favor the method used in the AAU study, and columnists who believe that the discipline is wrongheaded. Nblund (talk) 18:36, 19 January 2016 (UTC)

Your comment about Yoffe is completely wrong. In the article where she comments on the AAU study, she does not mention the Clery Act reports, but refers to the Federal Bureau of Justice Statistics findings which uses criminal definitions of sexual assault, "This is illustrated dramatically by the release last December of a special report from the Bureau of Justice Statistics...The report found that among women aged 18 to 24, those not in college were 1.2 times more likely to be victims of sexual violence than those in college. The good news was that incidence for both groups was far lower than anything approaching 1 in 5: 0.76 percent for nonstudents and 0.61 percent for students." [13]. You're pretty brazen with this misrepresentation given how easy it is to check; The Mother Jones article you linked to doesn't even mention Yoffe or the Bureau of Justice Statistics. The BJS report is based on an annually recurring, random phone survey which gathers the responses of victims - not what schools disclose under the Clery Act. Yoffe then goes on to comment on what the AAU survey stated - again not what the Clery Act reports disclose. Mattnad (talk) 20:45, 19 January 2016 (UTC)

Correction: Taylor, not Yoffe, is the one making an argument rooted in a misunderstanding of how the Clery Act works. Nblund (talk) 21:19, 19 January 2016 (UTC)

Ah! That does take a big bite out of Taylor and Yoffe's criticism. There was nothing to correct. @Mattnad: Look further down in Yoffe's article to where she writes " in a semi-annual report" - the link is to Clery Act data, and her subsequent argument is based on this. --Sammy1339 (talk) 21:35, 19 January 2016 (UTC)

Not really. The fundamental gist of Yoffe's observation is based on BJS data vs. AAU which demonstrates a chasm between what federal government finds using the criminal definition, and what the AAU study reported using very a very broad and subjective funnel. I'll add that the survey the BJS used for their finding found only 10.6% of women who didn't report their experience, did so because they didn't think it was serious enough. That's very different from the 60% to 75% in the AAU study. Yoffe's comments relating to Clery Data is just another data set but it doesn't go to the reason why women don't report. For that, she cites the AAU study. Mattnad (talk) 22:00, 19 January 2016 (UTC)

I'd like to add that the AAU itself did not qualify the dominant reason in its highlight summary. All they say is "More than 50 percent of the victims of even the most serious incidents (e.g., forced penetration) say they do not report the event because they do not consider it “serious enough.” If the AAU didn't think it was important enough to qualify it the way that Nblund would like us to, why should we?Mattnad (talk) 13:58, 21 January 2016 (UTC)

"The majority of rape and other sexual assault victims do not report their attacks to law enforcement."

I've gone ahead and expanded this discussion to WP:RfC input. For those viewing this from the WP:RfC page, my and others' arguments are below on the talk page. The RfC concerns whether or not to add WP:In-text attribution to the "majority of rape and other sexual assault victims do not report their attacks to law enforcement" aspect of the article and/or additional information for further context. Flyer22 Reborn (talk) 15:44, 17 January 2016 (UTC)

_____

Scoundr3l took issue with the following sentence: "The majority of rape and other sexual assault victims do not report their attacks to law enforcement." He added an "according to whom?" tag. I reverted, stating, "Read Template:According to whom. Also, we don't add in-text attribution in a way that can mislead; see WP:In-text attribution. [...] The source is right there at the end. Adding 'so an[d] so stated this' can make it seem like some sole person's opinion." I tweaked the text and duplicated the references via WP:REFNAME so that the material was left in this state. As this link shows, Scoundr3l then added "some" in place of "majority," stating, "That is precisely that point, I'm afraid. Since it's impossible to know how many of a thing are unknown, to say it is a majority is an opinion and I would recommend an in-text attribution in lieu of this much softer wording. Who said it was a majority?", and I replied. "We go by the WP:Reliable sources, not personal opinion. And, per WP:Due weight, we don't need to stack many sources to validate it." I then compromised by adding, "The literature indicates that" in front of "the majority." This also was not enough for Scoundr3l, and he added Template:Specify to "the literature indicates" part, stating, "We've come full circle. There is no reason this statement can not or should not be attributed directly to its source This is standard practice throughout this and similar articles and conforms to all your linked policies, as I'm sure you know."

Scoundr3l is wrong; we do not directly attribute something (meaning with in-text attribution) that is widely supported by the literature to one or a few people; WP:In-text attribution is very clear about that. And that most rapes go unreported is widely supported by the literature. That most sexual assaults go unreported is widely supported by the literature. There is no need to specify what is meant by "the literature"; anyone with common sense should know that we mean "the rape and other sexual assault literature." I just saw that Nblund also reverted Scoundr3l, stating, "This is consistent across several decades of research, attributing it to a specific group seems unnecessary, because it's a consensus that isn't really questioned. We can discuss in talk." Flyer22 Reborn (talk) 01:12, 15 January 2016 (UTC)

Yeah. Scoundr3l, to answer the question about how these figures are derived: these estimates are derived from surveys that ask people about criminal and other victimization, and that ask victims whether they reported the crime to the police. Surveys have a margin of error, of course, and there are reasonable disagreements about things like wording and survey context, but the finding that most sexual assaults are not reported is something that is consistent across -- quite literally -- every single data source discussed in the entry, and really every data source I'm aware of. I don't know of any experts who seriously question this finding. Where there's a consensus in a field, it's really not necessary (or possible) include attribution to any particular individual. Indeed, it can give the false impression of a disagreement where none exists (see WP:INTEXT). Nblund (talk) 01:27, 15 January 2016‎ (UTC)

Indeed surveys have a margin of error. They also do not fact-check, which is precisely why survey results are never presented as facts, but as survey results. Even the worst journalists know better than to present survey results without attributing them in-text because no amount of gallop polls will make "Most people's favorite color is blue" a fact. If I had access to the primary source, I'm sure even it would agree with me: it's a statistic, not an objective conclusion. "According to polls, 68% of people say their favorite color is blue", on the other hand, would be considered a fact (assuming there was a source to verify) and as an objective fact, it could be presented in Wikipedia's voice. If we can't rely on common sense, I'll instead point out Wikipedia:When_to_cite specifically suggests "Opinions, data and statistics, and statements based on someone's scientific work should be cited and attributed to their authors in the text." and you'll find this practice throughout every survey result in this article and other articles that rely heavily on conflicting surveys. If I'm not mistaken, though, Flyer22's motivation for reverting my request for attribution was not based on his assertion that it's an objective conclusion, but that he did not want it presented as a minority or sole opinion, which I don't disagree with. But I disagree with presenting second-party survey comments as facts in Wikipedia's voice. I don't even dispute the truth of it, only the wording. I know of several surveys that show numbers between 68% and 90% unreported, all we need to do is attribute them. Again, I don't have access to the original source, so I don't know what survey or surveys it's citing, hence the tag. If anyone has that information, or would like to provide other sources, perhaps we can agree on something like this "According to a numerous surveys, such as the National Department of Justice and the Bureau of Crime Statistics, most rapes go unreported", as an example. Or "between x% and y%" if we're going for accuracy. 2001:57A:400B:101:3409:7F75:9DFC:8974 (talk) 17:32, 15 January 2016 (UTC)

preceding IP is me, btw. Scoundr3l (talk) 17:33, 15 January 2016 (UTC)

Here's the wording from Unsafe in the Ivory Tower (the first cited source): Despite the prevalence of sexual offenses committed against college women, the majority of women who are victimized do not report the incident to the police (Fisher, Daigle, Cullen, & Turner, 2003a). There's parenthetical attribution, but it's stated as a fact in the text. The book goes on to cite seven other studies with this finding, and that's not an exhaustive list. Citing seven surveys, in-text, seems excessive, and it's probably not necessary.

Peer reviewed surveys generally do involve fact checking, critique, and replication, and I don't think you're correct that Wikipedia articles always provide in-text attribution for statistics. The entries for diseases, for instance, usually provide an estimate of the prevalence of a disease without in-text attribution, because these numbers are basically uncontroversial.

I think the "When to Cite" guideline is referencing cases where a specific statistic is attributable to a particular author -- for instance -- we attribute the finding about repeat offenders to David Lisak, because that is a finding that is unique to his work. In this case, however, we're citing a tertiary source that makes a general statement about findings of multiple authors.

I think that sentence is un-problematic, but I would be open to including subsequent sentences that bolstered the statement by pointing to specific sources for this finding. Nblund (talk) 19:02, 15 January 2016 (UTC)

Thank you for isolating the primary source. Here are the end conclusions as taken directly from that source "Acknowledging sexual victimization as rape: Results from a national-level study." Fisher, Daigle, Cullen, & Turner (2003) Justice Quarterly:

*'Our findings, though not definitive, have clear implications for this debate. By building on the best measurement strategies available-in essence, synthesizing the work of researchers like Kilpa [...] we have attempted to develop a measure of rape that is less susceptible to the criticism that we are counting as rape acts that really do not qualify under this legal category'

*'Thus, using a different measurement approach, our findings lend support to the proposition of Koss and other researchers that the acknowledgment of rape is a real problem, not a methodological artifact. We found that only about half the rapes were acknowledged by the victims. Our estimate, as noted previously, is about twice as high as that reported by Koss'

*'In any event, our study suggests that over half the women in the sample who were raped did not acknowledge this event as a rape'

Implications, supports, estimates, and suggestions. The writers of the report at least acknowledge the difference between a datum and an objective truth. You'll find the same thing in any reliable study, I'm sure. Survey results are not facts and second parties who draw conclusions from the data should be attributed as it is their opinion and often not the conclusion present in the primary source. I hope the proposed compromise helps establish that this is a majority finding across all major studies (at least as far as we've seen), but still attributable to those studies and not Wikipedia. That should hopefully make everybody happy. Hopefully. Scoundr3l (talk) 19:55, 15 January 2016 (UTC)

The entry cites Fisher's 2009 book, which itself cites Fisher et. al 2003. The quotes you're providing are actually a different 2003 publication by Fisher et al. from the one cited in the source, and those statements don't reference findings about non-reporting to the police. Here's the referenced article. Here's the relevant quote from that source.

Quote

Despite the prevalence of sexual offenses, a large proportion of victims did not report their sexual victimization to the police or to other authorities (Tjaden & Thoennes, 2000). Results from the National Crime Victimization Survey (NCVS) have consistently shown that rape and sexual assault have been the most widely underreported violent crimes. In fact, the 1999 NCVS results revealed that only 28.3% of these crimes were reported to the police (Rennison, 1999). Notably, other research has provided even lower estimates of reporting (Bachman, 1998; Finkelson & Oswalt, 1995; Gartner & Macmillan, 1995; Kilpatrick, Edmunds, & Seymour, 1992; Kilpatrick, Saunders, Veronen, Best, & Von, 1987; Koss, 1985; Russell, 1983; Tjaden & Thoennes, 2000). Similarly, sexual victimizations of college women have gone largely unreported. To illustrate, Koss et al. (1987) found that only 5% of college student rape victims reported their experience to the police. In a national study of college students, Sloan, Fisher, and Cullen (1997) found that only 22% of rapes and 17% of sexual assaults were disclosed to local police, county sheriff, campus police, campus security, or other authorities

This formulation is similar to what I proposed: it begins with a general conclusion about the prevalence of non-reporting, then cites specific supporting evidence. This seems like a reasonable way to address your concern. If your proposed standards were applied, it seems like virtually every bit of statistical information would require in-text attribution. This clearly isn't the case: the entries for unemployment in the U.S., voter turnout, the epidemiology of Parkinson's disease are all rely on statistical estimates, and -- in every entry -- statistics are cited without in-text attribution. Are these problematic as well? Nblund (talk) 23:25, 15 January 2016 (UTC)

Well, once again, I do not have access to the source, which is why I requested attribution. With only the citation to work with, I had to assume what quotes are being referenced. As for the new quote, I hope you've noticed that every claim in this quote attributes it either directly to the source or to "other surveys" which it then cites. Only the vague claims like "a large proportion" are not directly supported in-text. I would be fine with changing the text in this article to "a large proportion", as an example, but I'd rather reflect a more accurate number and where the information came from. As with your previous example, the obvious difference is that HIV epidemiology, voter turnout, etc are objectively measurable. You don't survey people and ask if they have HIV or if they voted, you measure quantifiable data. That is not the case with these surveys, so their conclusions are not "x amount of people were affected by z" it's "x amount of people reported being affected by z". And since we're not here to call them liars, or debate onus probandi, what we have to do instead is accurately report the information as it was received. I'd be interested in seeing an example of your proposition because I don't think we're disagreeing as much as it may seem like we are. Scoundr3l (talk) 00:11, 16 January 2016 (UTC)

Totally understandable, the first quote is from the source cited in the entry. The second quote is from the source cited by the source cited in the entry. Both quotes provide citations, but they also both make general statements of fact about the prevalence of unreported rape without providing in-text attribution. The first citation uses "a majority" without in-text attribution, so it seems like this is a perfectly reasonable way to do it in the entry. We could bolster that argument by citing specific studies in subsequent sentences.

BLS unemployment data come from a monthly household survey. HIV/AIDS data comes from asking health departments how many people they diagnosed. Those numbers don't necessarily count all diagnoses, and are usually adjusted to account for under-reporting. I'm not sure I follow the distinction you're drawing between self-report surveys and other kinds of data, but it seems like other entries attribute widely accepted survey data any differently from any other scientific fact. Nblund (talk) 00:52, 16 January 2016 (UTC)

Which part of the distinction did you not follow, exactly? A self-reported survey is an unsupported claim, while other forms of data are objectively verifiable. If a conclusion is not based on objective data, it's not a fact, it's a proposition or an estimate. Additionally rape is a crime and all crimes, at least in my country, are subject to due process. The accused are innocent until proven guilty in court of law, so any unreported crime is inherently an unverified claim. I can't imagine any of that's what confused you, so are you disputing that it's based on an estimate or merely that estimates don't require attribution? If we're using examples from other articles, I'd be happy to find 8 examples of attributed estimates for every unattributed example, but perhaps we can save time if you'd just explain why this statement should not be clarified. Aside from 'I don't think it's necessary', have you got any reason to dispute me clarifying where this information came from, assuming of course that it's objective and balanced? I have plenty of sources I can use if you aren't willing to use your's, I just need to know if and why good information will be reverted.Scoundr3l (talk) 06:44, 16 January 2016 (UTC)

Scoundr3l, I disagree with you per what I stated above. You stated, "Flyer22's motivation for reverting my request for attribution was not based on his assertion that it's an objective conclusion, but that he did not want it presented as a minority or sole opinion, which I don't disagree with. But I disagree with presenting second-party survey comments as facts in Wikipedia's voice." Well, just to get this out of the way: I'm female. As for your points, not only do I not want the "most rapes and other sexual assaults go unreported" data presented as a minority or sole opinion, which is indeed what your in-text requests would have resulted in, I do not want it to seem like it's something that is less supported in the literature than it is. I repeat: It is widely supported. It is consensus. There is absolutely no reason at all to alter the text to name a few researchers or a few organizations, when many researchers and organizations state the same thing. Above, I pointed you to Google Books sources stating the same thing. Many WP:Secondary sources state the matter as fact. And so should Wikipedia. Other than that first compromise I gave you, I will not compromise on this. We can take it to a WP:RfC, but I won't be changing my mind. Flyer22 Reborn (talk) 20:45, 15 January 2016 (UTC)

On a more personal note: There is the fact that I've know many teenage girls and women throughout my life who were sexually assaulted in public in broad daylight and never reported it; by this, I mean a breast grab, a slap to the butt, and so on. And I don't mean by boyfriends. I mainly mean that this was done by male acquaintances who were not at all romantic with the girls or women, or was done by male strangers. I can't say how many times I've seen this happen with my own eyes. The girl or woman usually just hits the guy with a "jerk!" type of face, calls him a name, or laughs it off. I have no doubt that many other women have seen the same. So not only do I believe the data that most sexual assaults go unreported, I have personally seen it in life. One fault with the data is that it mainly focuses on female victims, but that's because girls and women are documented as the majority of victims of rape and other sexual assaults. For machismo and other cultural reasons, it also seems that males are less likely than females to report being victims of sexual assault or rape. Flyer22 Reborn (talk) 21:39, 15 January 2016 (UTC)

Do what you need to do. My requested changes do not have any effect on how well supported the statement is, it merely attributes the statement to its source rather than presenting it in Wikipedia's voice. Wikipedia isn't the one drawing that conclusion, the source is drawing it from survey data, so it's the source's conclusion. Your change to "the literature" was a step in the right direction, but unfortunately I find it too vague and unencyclopedic to be useful to the reader. Surely "the literature" doesn't all share this conclusion, and it's been demonstrated that it does not (Fisher, et al. does not present this conclusion). Instead, let's attribute it to what it is: survey results. We don't need to list every survey to do that, only where the conclusion is coming from. So why then would you not accept "multiple surveys", "numerous surveys", or even "general consensus" in lieu of "the literature"? Admittedly, we should source the consensus (per WP:RS/AC) but it's demonstrable and a step up from attributing it to divine providence. Also, I apologize for assuming you were a 'he'. Scoundr3l (talk) 21:42, 15 January 2016 (UTC)

I appreciate your honesty, but for obvious reasons, we should try to set our personal feelings aside for this issue. It's not my intention to diminish the validity of the statement, so let me know if you think of any way we can attribute it without doing so. Scoundr3l (talk) 21:46, 15 January 2016 (UTC)

We agree to disagree then. I would not accept "multiple surveys" because "multiple" is often interpreted as "three or more" or "a few." I also don't think we should state "surveys" unless that is directly supported by the sources. "Numerous surveys" is WP:Weasel wording and is unnecessary. I could accept "general consensus," but it's not needed.

And apology accepted. It's well known that Wikipedia is mostly made up of males; so I expect others to assume that I'm male unless otherwise stated.

As for personal feelings, I don't edit with my personal feelings (well, not unless I'm in some heated dispute where emotions tend to spill over); my user page and talk page are clear about how I edit. I was simply offering you a perspective that you likely have not witnessed, at least to the same degree as females. Flyer22 Reborn (talk) 21:54, 15 January 2016 (UTC)

'Numerous surveys' was shorthand for the original demonstration above, which provided examples. This original example was more than enough to satisfy WP:WEASEL and any other slippery slopes of citation overkill.Scoundr3l (talk) 22:04, 15 January 2016 (UTC)

"Numerous studies," "many studies" and similar (yes, things like "numerous surveys" too) are sometimes needed, but I've also seen some Wikipedians argue (or seem to argue) that the wording is weasel wording or close to it; so I prefer to avoid it when I can (as you know, I believe this is one of those "can" cases). Flyer22 Reborn (talk) 23:37, 15 January 2016 (UTC)

Seeing as the technical name for weasel words is "unsupported attributions" and I'm the one in this discussion on the side of an in-text attribution, I don't think weasel words will be an issue on this end. Any of the above examples are simply more specific synonyms of literature. Scoundr3l (talk) 06:44, 16 January 2016 (UTC)

"Numerous" has been argued as vague, just like "many" has been argued as vague, and "vague" is one of the primary things that the WP:Weasel word guideline addresses. But like I noted, such terms are sometimes needed; not in this case. I don't see how "multiple surveys" or "numerous surveys" are specific synonyms of the literature in this case. As made clear by Nblund and me above, the literature usually states this matter as fact; this is because it's rarely disputed. WP:Due weight allows us to simply state the matter as fact. Flyer22 Reborn (talk) 17:47, 16 January 2016 (UTC)

I'm sure you'll remember that my original suggestion was a direct attribution, so factor into your calculations that anything vague or resembling weasel words is an attempt to comprise with your refusal to attribute the statement. In other words, unless you have an alternative suggestion, citing what constitutes a weasel word is both ironic and unnecessary. Additionally, by refusing to attribute the statement, you are interfering with the ability to provide the additional information you deem required to attribute the statement. I can not elaborate on a statement if I don't know what statement is being used. Due weight only comes into play if I were attempting to promote a minority viewpoint when in fact I am attempting to provide additional information on the cited majority viewpoint. Kindly assume good faith in that regard as I am assuming you are not attempting to hide the original source of information behind vague but verifiable claims. Aside from undue weight, what other objections might you have for providing this additional information? Scoundr3l (talk) 23:26, 16 January 2016 (UTC)

I'm not sure what you mean; the statement is already attributed, just not directly attributed to any one person or group of people. There is no need for a "direct" (meaning in-text) attribution. I've already made my case against in-text attribution above and against other suggestions you've made. So has Nblund. To state more would be repeating myself. WP:Due weight very much applies in this case. Nowhere have I stated anything about your lack of good faith. And my refusal to compromise further on something that does not need to be compromised on is not about any lack of good faith on my part; it's about what I've already stated on the matter. Flyer22 Reborn (talk) 01:12, 17 January 2016 (UTC)

The only reason you've provided so far is that in-text attribution would provide undue weight to the idea that this is a minority opinion. Nowhere does this policy allow you to prevent the addition of source information because that information might somehow lessen the strength of those claims. It's either supported as strongly as you claim or it's not. Providing evidence of the support couldn't unduly take weight away from the claim, unless of course the article text is inaccurate. Or unless you're assuming it'll be done unfairly, which you've no right to assume. If providing the source in any way weakens the statement, this is a failing of the source and not of Wikipedia in reporting on that source, nor is there any legitimate reason to prevent the addition of that source under due weight. IF your stance is that there is an academic consensus, by all means, find a source other than a Google search to support that there is a consensus. That'd be a welcome addition to the article because it would at least provide some measure of objectivity to the claim. What we can't do, however, it conceal where the information comes from in order to actively protect a viewpoint, majority or otherwise. Yes, there is a citation of the comment, but not to what it's referring. By concealing that information, you prevent the addition of other commentary and details from the source. The facts must support themselves and the readers and editors need to be provided with enough information to verify that. Any personal thoughts you have on whether the change is necessary are not under consideration, but just to be polite, I'll share one simple and valid reason I need the change: I want to know the exact figure. I can't do that with vague commentary. I also feel the sentence should be written in a way that is not subject to cultural bias or data rot. Did you know that the Fisher & Daigle report was US only? Because this article doesn't. I can't correct it until the text reflects what's being cited. By all means, don't repeat yourself, but if you have a reason I have not yet heard, please let me know; although we should probably break line as it's getting cramped over here. Scoundr3l (talk) 06:39, 17 January 2016 (UTC)

With regard to your "IF your stance is that there is an academic consensus, by all means, find a source other than a Google search to support that there is a consensus.", statement, WP:Due weight is about what the preponderance of sources state; I linked to two Google Books links showing that "the majority of rapes and other sexual assaults go unreported" aspect is widely supported. That wide support is also why the lead of the Rape by gender article currently states, "Since only a small percentage of acts of sexual violence are brought to the attention of the authorities, it is difficult to compile accurate rape statistics.", with two sources supporting it. Unless you can show that the preponderance of sources state otherwise on that, I've satisfied my WP:Due weight argument in that regard. And the "the majority of rapes and other sexual assaults go unreported" aspect is not at all limited to the United States. You stated, "I want to know the exact figure." There is no one figure, but the figure is generally that "the majority of rapes and other sexual assaults go unreported." I know how WP:Due weight works and how WP:In-text attribution is supposed to work, and that's what I've been arguing above. I'll only accept additional information for further context if done right. You can assume my thought processes (such as the incorrect assumption of "concealing information"), but unless those thought processes are made perfectly clear, it's better not to assume. My discussion with you on this has pretty much ended. I've now turned it into a WP:RfC for further input. Flyer22 Reborn (talk) 15:44, 17 January 2016 (UTC)

You seem to be merging two of my points, so I apologize if I was unclear. I've made no argument that it's not a majority viewpoint. As such, your Google search results are not germane, nor admissible as evidence of a anything related to my position. 5 million search results is not the same thing as 5 million sources, as I'm sure you know, but it's irrelevant to the conversation. Above, you directly claimed there was a consensus (which I don't quite equate with a majority viewpoint) to which I assume you meant an academic consensus as defined by Wikipedia policy. You'll find your statement above and in bold. Since you seemed quite sure of it, I asked you to find a source since proof of a consensus would be a happy resolution for both of us. It seems by consensus you meant your Google results, which you may consider proof of consensus, but unfortunately Wikipedia does not. Otherwise, it appears you're not pursuing the consensus argument any more, so we'll table it. It was a secondary point, anyway. That leaves us with majority viewpoint, to which I've made no argument, nor do I have opinion, so I have no reason to pursue that counter argument. For this discussion, let's go ahead and agree that it's a majority viewpoint. So instead, when I'm asking is how my changes are promoting a minority viewpoint. My proposed changes are only to bring more information in from the existing citation, which supports this majority viewpoint. Since you are sticking with Due Weight, I invited you to explain how providing additional information from these sources promotes a minority viewpoint. You've not yet answered that question and it's the initial reason you've provided for contesting my changes. Due weight is not a license to prevent any changes you deem may be harmful to a position, so please explain what part of that policy warrants the prevention of an in-text attribution. As you've hopefully re-read WP:INTEXT during the course of this discussion, you'll remember that it takes no explicit stance on when you must or must not include an in-text attribution, so your opinion on how that policy is supposed to work is hearsay. And since you're the one who initially took issue with my edits, there is generally a greater onus on you to explain your reverts than I to explain my bold edits. You've linked policies which don't seem to support your revert and been challenged to support them. This is why we have talk pages. Your declarations that you're done with the conversation are understandable but not to be confused with a conclusion. Nobody can make you engage in this discussion, but neither are your insights necessary to finding a resolution. Scoundr3l (talk) 00:14, 18 January 2016 (UTC)

My Google Books argument was germane, per what I initially stated at the beginning of this section. You added tags, then wording, to the sentence at hand in a way that was inappropriate; why they were inappropriate were noted above.

Anyway, like I noted, unless I am discussing proposed wording with you below, I don't have anything more to state to you about this, especially since you keep assuming what I meant in instances when it's not what I meant. In my opinion, I can't be any clearer than I've been. You and I have different understandings of how Wikipedia works in this case. For example, your assertion that WP:INTEXT "takes no explicit stance on when you must or must not include an in-text attribution [and so my] opinion on how that policy is supposed to work is hearsay" is incorrect; that guideline (not policy) is very clear that the type of in-text attribution you were initially asking for would have been misleading. If anyone not familiar with the literature had answered your tag, it would have resulted in a clear-cut WP:INTEXT violation. You even added "some" in place of "majority," which clearly undermined the literature. You and I also have a different definition of WP:ONUS. The onus is on you to convince others why we should not report the "majority of rapes and other sexual assaults go unreported" aspect as the fact that it is. WP:ONUS states, "The onus to achieve consensus for inclusion is on those seeking to include disputed content." And if I were stating that this discussion is done because I am about done talking with you about this matter, I would not have started a WP:RfC. Flyer22 Reborn (talk) 01:02, 18 January 2016 (UTC)

I was using the traditional, dictionary definition of onus, from Latin meaning "responsibility". As in, "its the responsibility of the reverter to show and prove the consensus in question" as taken from the WP:DRNC page, in the context of words as they may sometimes be used outside of Wikipedia articles. This onus would logically precede your above onus, despite both being irrelevant to a discussion on policy, not consensus. Still, while you were able to quote that article, what we really need are your quotes from WP:INTEXT or WP:Due weight which support your position to exclude this information. If we focused more on that, I think the discussion would have moved along a lot faster. Scoundr3l (talk) 05:42, 18 January 2016 (UTC)

*Please note We are current discussing the formatting of this template on Flyer22's talk page. You may wish to hold your comments until the resolution of that discussion.Scoundr3l (talk) 00:42, 18 January 2016 (UTC) Strike-through mine. Scoundr3l (talk) 04:50, 20 January 2016 (UTC)

I'll respond to the RfC as phrased about the line stated, but feel I should offer the view that the wording of the line creates an self-defeating impression of unsupported soapboxing or an internet-exaggeration, and the mentioned 'numerous studies' might at least have avoided that impression and the whole in-text RfC. Markbassett (talk) 16:27, 22 January 2016 (UTC)

Don't use in-text attribution

• Support, per my and Nblund's comments above. Flyer22 Reborn (talk) 15:44, 17 January 2016 (UTC)
• Support per Flyer22 Reborn and Nblund. Kaldari (talk) 00:14, 18 January 2016 (UTC)
• Support per Flyer22 Reborn and Nblund. Neljack (talk) 12:12, 18 January 2016 (UTC)
• Oppose - why not disclose the sources in text? So the reader cannot as easily understand where it's coming from. That does a disservice to the reader. Mattnad (talk) 14:33, 18 January 2016 (UTC)
• Objection - The framing of this RtC is unusual. I don't see the need for two opposing areas of comment to answer one question. To prevent redundancy and over-complication, I will post my comment to the simplified question Should this statement include in-text attribution? only in the section immediately below. Scoundr3l (talk) 20:47, 18 January 2016 (UTC)

I responded on my talk page to Scoundr3l, noting that the framing is not unusual. Flyer22 Reborn (talk) 00:23, 19 January 2016 (UTC)

Unfortunately, the discussion was slow and fruitless, so I've noted my objections here in order to proceed. There is nothing usual about this framing of an RfC, hence it is unusual, and also wholly unnecessary. That you can find other examples of unusual framing do not explain why it was done this way or explain why this framing is necessary. The redundancy and over-complexity should be self-evident. Rather than a simple list of yays and nays, we now have a partial list of yays and nays to the affirmative and a partial list of yays and nays to the negative, some who've commented on both. If you would like further information on how to RfCs are usually framed, there are examples on that page. Since it's too late to be corrected, I only hope you'll consider this in your future attempts at framing RfC. Scoundr3l (talk) 04:32, 20 January 2016 (UTC)

I'm not going to debate this matter with you too. I was clear that I disagree with you and why. Flyer22 Reborn (talk) 21:01, 20 January 2016 (UTC)

Note that I also considered redundancy, but felt that the setup would be fine. My voting twice doesn't mean others have to as well. I also considered voting once. I voted twice because it is common to do so in RfCs. From what I see, this RfC is largely fine, along with the "Proposed wording" and "Further commentary" sections developing in the way were designed to develop. Flyer22 Reborn (talk) 21:10, 20 January 2016 (UTC)

Your disagreement has been noted, but the collaborative efforts of Wikipedia do not rest for whether or not Flyer22 agrees with them. If you're not inclined to debate, perhaps you should take less stock in what you personally agree with and focus on things everybody can agree with. Scoundr3l (talk) 00:50, 21 January 2016 (UTC)

And there you go again presuming what I think or don't think. On Wikipedia, despite its collaborative nature, no one is obligated to debate something they don't want to debate, just like they are not obligated to do so in real life. Even so, I indulged you on the heading matter. Then it was over because the matter is no big deal. It would be wise for you take your own advice when it comes to disagreement, since it is your disagreement with two editors who are clearly more familiar with the literature on rape and sexual assault than you are that led to all of this. Flyer22 Reborn (talk) 03:22, 21 January 2016 (UTC)

But since I don't want to keep arguing with you about this, I struck my "03:22, 21 January 2016 (UTC)" post. Flyer22 Reborn (talk) 04:19, 21 January 2016 (UTC)

I appreciate it. Presuming my level of knowledge with the literature and whether or not you're more qualified to discuss is not only absurd, it flies in the face of the yet unswayed consensus. I don't presume to know your thoughts, but speaking for myself: I'm not here to prove you or anyone else wrong, I'm here to improve the article. Scoundr3l (talk) 04:40, 21 January 2016 (UTC)

You noted on this talk page your lack of familiarity with the literature on this particular aspect we are debating. And when you make comments about me like "the collaborative efforts of Wikipedia do not rest for whether or not Flyer22 agrees with them," you are categorizing my mindset based on your own assumptions; stop it. Stop taking disagreement and/or an editor's right to not want to debate you on something, especially when that editor is trying to avoid arguing in circles, so personally. Flyer22 Reborn (talk) 05:00, 21 January 2016 (UTC)

And when you make the kind of edits you made to the text in question, such as adding "some" in place of "majority," coming to the conclusion that you are not as familiar with the literature is understandable. Flyer22 Reborn (talk) 05:08, 21 January 2016 (UTC)

I never said I was unfamiliar with the literature, I said that I don't have access to the source in order to find what it's quoting. Furthermore, knowledge of the subject has little to do with this discussion. Whether or not a statement is subjective enough to be attributed is a matter of a priori logic and my argument in this case is Socratic. Replacing "some" with "majority" was an attempt to compromise your refusal to attribute the claim. This was also based on simple logic: a less sensational claim requires less sensational support. If you equate this with unfamiliarity with the literature, that's your own bias. Likewise, that the discussion does not rest pending your agreement is simply an objective fact in response to your repeated declarations that you are done with a discussion. I'm not asking you to engage in the conversation or debate me, I'm simply informing you that a refusal to compromise will not be seen as an obstacle to resolving the issue.Scoundr3l (talk) 08:18, 21 January 2016 (UTC)

Knowledge of the literature has much to do with this discussion, and that is displayed in spades across this talk page. Replacing "some" with "majority" was undermining the literature, pure and simple. No one familiar with the literature would have added "some" in place of "majority." The claim was already attributed. You were asking for inappropriate in-text attribution with your Template:According to whom tag. I've been clear above that such tagging would have resulted in inappropriate in-text attribution had someone unfamiliar with the literature answered that tag's request. It would have made that sentence seem like the matter is only according to that person or group of people. I never stated that I was done with the discussion; I stated that I was about done discussing things with you. You know, to avoid the type of circular argument we are having right now (I easily recognize when a discussion with a certain editor is going to be circular; I'd rather not continue to discuss matters with those type of editors). I then expanded the discussion to a WP:RfC, which clearly was not me stating that the discussion is over. If I or someone else does not want to discuss things with you any further, you should learn to accept it instead of resorting to all sorts of "you aren't playing fair" or "you are being difficult" claims, or other misguided assumptions. You love to debate, everything apparently. You couldn't even drop this debate after I struck through my above comment; removing my comment probably would not have even stopped you. An editor does not have to compromise on something they feel needs no compromising on. I tried compromising with you; that is clear from the very beginning of this thread; that was not enough for you. And now we are here. Below, I was clear about the only compromise I will accept on this matter. Flyer22 Reborn (talk) 21:24, 21 January 2016 (UTC)

That you continue to share your opinions on other matters, most of which are impertinent to the conversation, is the only reason we're still in discussion. I have every right to reply to your comments, especially when they've devolved into petty personal remarks, but the only person who can make you continue the discussion is you. Since you've not yet established a concrete case for exclusion, perhaps you've over-estimated your level of expertise on the matter or your ability to read other editors. It may be more constructive to contribute some additional rationale to the discussion, as merit alone has done little to impress support. Also, I think you may be quoting somebody else. I never made either of those claims. If you're still unwilling to further compromise, that's your choice, but ultimately the decision isn't up to you, so whether or not you're 'playing fair' or being difficult are not of concern to me. As for your refusal to compromise on the framing of an RfC, I've never seen that before, so I can only express bewilderment. Scoundr3l (talk) 04:33, 22 January 2016 (UTC)

Wrong about all of that, with the exception that you have the right to reply. Now we'll see if you will debate "Wrong about all of that." too. The petty, personal remarks initially came from you, which is why I made my "05:00, 21 January 2016 (UTC)" and "21:24, 21 January 2016 (UTC)" comments above. The "you aren't playing fair" or "you are being difficult" claims are the type of claims you've made in this discussion (and do spare me any "well, if you are seeing that, it must be true of your behavior" remarks), but nowhere did I state you made those exact comments. As for "little to impress support," that's your opinion. WP:RfCs commonly go slow these days, and the participation in this one is small so far. I'm not the one leaving notes in the WP:RfC in an attempt to influence it, and I'm not the one contacting an editor to weigh in. Flyer22 Reborn (talk) 18:49, 22 January 2016 (UTC)

I think you'll find that leaving notes in an RfC is, by definition, the purpose of an RfC. Of the two of us, only one of us has challenged another commentator's comment. You're also the only one who keeps bringing up right and wrong, when it's clearly a difference of perspective. Your personal editing style and definition of civility is your business, but as you are a clearly a non-expert in the field of my opinion, your attempts at interpreting and paraphrasing my claims are to be treated as hearsay at best and petty stone-throwing at worst. I have indeed contacted many users, including those who support your claims, in order to improve participation in this page. Nblund has as well. Kudos on the detective work. If you've got comments that are related to the discussion, though, perhaps we should focus on those. Scoundr3l (talk) 20:01, 22 January 2016 (UTC)

You stated, "I think you'll find that leaving notes in an RfC is, by definition, the purpose of an RfC." Not your kind of notes. And you've made more incorrect characterizations of me in your "20:01, 22 January 2016" post, as expected, I see. As for expertise, I've been quite clear about which of us is more familiar with the literature; my reputation on these topics (sexual topics, including sex offenses) speaks for itself; there is no need for me to justify myself to you. That you take my lack of interest in indulging you as a sign of lack of knowledge is silly, but it's clearly your right to think what you want. Yes, I saw that you contacted others; the others were contacted about the proposed wording section, and they are editors who already weighed in. They were contacted after you contacted that one editor to weigh in. You stated, "[my] attempts at interpreting and paraphrasing [your] claims are to be treated as hearsay at best and petty stone-throwing at worst." Yes, that is exactly how I feel about your claims toward me. Now if you are done with this particular debate, I will be WP:Hatting this bickering. Or is it that you want the WP:Last word, as you clearly seem to always do? Flyer22 Reborn (talk) 20:46, 22 January 2016 (UTC)

It's Friday, both of our objections have been noted, and this thread has officially derailed. You misunderstood my comment about your expertise, I wasn't saying "in my opinion, you are not an expert" I meant "you are not an expert in matters relating to my opinion". But rather than continue to fight over semantics and baseless interpretations of character, can we agree to call it quits and work on the relevant discussion? Scoundr3l (talk) 20:50, 22 January 2016 (UTC)

Seems like we're in agreement, though in the interest of your 'last word' comment, I'll point out that we wrote our above comments semi-simultaneously. Scoundr3l (talk) 20:55, 22 January 2016 (UTC)

WP:Hatted. Flyer22 Reborn (talk) 20:59, 22 January 2016 (UTC)

Because of this and this, I have unhatted the discussion. I do not have the time or patience to debate every single thing with you. So fine; leave all the bickering on open display. Flyer22 Reborn (talk) 21:12, 22 January 2016 (UTC)

• Oppose per my comment below. More information is better here. --Sammy1339 (talk) 01:05, 19 January 2016 (UTC)
• Support Nblund (talk) 21:40, 19 January 2016 (UTC)
• Oppose meaning 'show the cite(s)'. The RfC is asking about the line shown, which currently shows no reason to give it credence and for WP does not convey the factual phrase origin. Just follow the cites, convey with fidelity what is out there and show where it is from. A desire to avoid citing because it is from a single source indicates to me a problem with being honest about that, and also that the phrasing is generally NOT said in a relevant community consensus statement nor common among reputable sources. What is shown also conveys the impression of soapboxing, so seems pointless anyway, as well as bing one view above due WP:WEIGHT. Markbassett (talk) 16:56, 22 January 2016 (UTC)

Use in-text attribution

Note: There is no obligation to vote twice. Feel free to ignore voting in this section. Flyer22 Reborn (talk) 20:52, 22 January 2016 (UTC)

• Oppose, per my and Nblund's comments above. Flyer22 Reborn (talk) 15:44, 17 January 2016 (UTC)
• Support since it provides context to the reader.Mattnad (talk) 14:32, 18 January 2016 (UTC)
• Comment I would just like to clarify a few points on my position:
  • This is not a position on whether or not the statement is true.
  • This is not a position on whether or not the statement is cited and verifiable.
  • This is not a position on whether in-text attribution is required
  • This is only a position that there is a strong case for in-text attribution and a weak case against. -Scoundr3l (talk) 21:37, 18 January 2016 (UTC)
• Support My reasons for supporting in-text attribution are as follows:
  • The current wording is prohibitive to further improvement of the article

The figure is being drawn from a specific source or sources of data. By not including the data source in the text, additional commentary related to those data sources is impossible. In this way, the quoted source is given artificial sanctity over any commentary related to the data source, when it is surely not the only source of commentary.

• The current wording is based on survey data

There are a number of conflicting estimates for this data, but they are all estimates. Uncertain facts should not be written in Wikipedia's voice.

• The current wording is time-sensitive

As time goes on and things change (in this case, the sooner the better), this statement may be found to be objectively false. An attributed statement would always be objectively true.

• The current wording is limited to the availability of data

If even one source emerges which contradicts this conclusion, the statement will be contentious. An attributed statement will never be a matter of contention.

• The current wording does not conform with the rest of the article text

The entire structure of this article is about providing a study and then balanced commentary on that study. In my opinion, this article is a monument to balanced commentary on a controversial subject. My by count, there are 82 statistical claims in this article's body text, of which 63 (76.8%) are directly attributed to their data source in the text. 13 (15.8%) are attributed generically to a "study" or "survey" which is then cited. 5 (6%) are not attributed but provide specific data. Only 1 other sentence in this article body is not attributed to a source, provides no specific data, and provides no further information (i.e. 'a study') as to where it came from. Perhaps coincidentally, that other sentence shares a source with this sentence.

• Attributing a statement can not unduly take weight away from the statement

The statement must stand on its own. The source of a statement is a verifiable fact. If revealing where the data comes from has any negative impact on the statement, it is solely the failing of the source and not undue reporting by Wikipedia. Contrarily, concealing the source of data can unduly lend weight to a statement that it may otherwise not have. The source is the source, for better or worse.

Given those points, and the lack of good points for exclusion, I see no reason this statement should not be attributed to its data source. Scoundr3l (talk) 21:37, 18 January 2016 (UTC)

• Support. Even thought the statement is indeed widely agreed-upon, it is helpful to the reader to have the context provided by the attribution, due to the wide divergence in operational definitions used in this subject. Better still would be a note showing that this statement is supported by multiple studies that used different operational definitions and methodology. --Sammy1339 (talk) 01:03, 19 January 2016 (UTC)
• Support, per Scoundr3l's reasons throughout. Stating survey results as fact is intentional misrepresentation and misleads readers. Statement should clearly be attributed to its source(s), and its origins and methodology discussed. This isn't scientific fact, as we're all aware human surveys elicit very specific behaviour responses, and cannot be treated as generalisations, no matter how many surveys are cited. FoCuS contribs; talk to me! 16:42, 19 January 2016 (UTC)

FoCuSandLeArN (FoCuS), how is it misleading the readers when the "majority of rape and other sexual assault victims do not report their attacks to law enforcement" aspect is widely supported in the literature and there is no indication whatsoever that it is false? How does it make sense to attribute that statement to one or more sources, as though it's only according to those sources that "the majority of rape and other sexual assault victims do not report their attacks to law enforcement", when it's actually supported throughout the literature, in scholarly books, legal documents, etc.? How is providing in-text attribution for this not a WP:INTEXT violation? Sammy1339's support vote above does not even seem to be support for in-text attribution, which would be misleading, but rather for more context. Appropriate context. We include the "most sexual assaults go unreported" or "most sexual violence is unreported" aspect in rape articles and related articles here at Wikipedia, including in the lead of the Rape by gender article. Are you suggesting that we give in-text attribution in all of these cases, or further context even in the lead, when the statement is not at all contentious in the literature or validly disputed in the literature? Flyer22 Reborn (talk) 18:24, 19 January 2016 (UTC)

I'm not judging the veracity of the literature, nor did I ever imply doing so. I'm saying it is undue to present it as a fact when in reality it's a poor statistic (exactly why surveys are rarely used as basis for policy), and as such it must be indicated. Surveys aren't scientific fact; they're sociological instruments subject to human behaviour and circumstance, reflecting opinion in response to often poorly-formulated questionnaires. I was summoned by the RfC bot, and provided my !vote after reading the above discussion. By the look of things, your attitude leaves much to wish for. Respect other people's views and let the RfC progress. Regards, FoCuS contribs; talk to me! 19:28, 19 January 2016 (UTC)

FoCuSandLeArN (FoCuS), asking you to explain your rationale for your vote is not disrespecting your vote or you. Since you've been involved in a number of RfCs, you should know that. Needless to state, I disagree with your rationale, and have given ample reason why I do. As for attitudes, I feel the same regarding you judging by your "19:28, 19 January 2016 (UTC)" response. Flyer22 Reborn (talk) 19:53, 19 January 2016 (UTC)

I provided my rationale the first time around. Deal with it. Your reputation precedes you. Best, FoCuS contribs; talk to me! 19:59, 19 January 2016 (UTC)

FoCuSandLeArN (FoCuS), indeed it does. Flyer22 Reborn (talk) 20:07, 19 January 2016 (UTC)

Let's try to stay focused, guys. So far, this has been pretty productive, but there is already a lengthy discussion above which addressed these questions and came to no resolution, hence the "per <user>" commentary. Beginning the discussion again is a step in the wrong direction. Obviously, there is disagreement over whether this statement stands on its own or needs additional information. All the talk-page evidence in the world has no effect on the body text, nor does it aid the reader, which is what we're here to discuss. Let's please keep our opinions of other editors and interpretation of their comments out of the discussion for now. Scoundr3l (talk) —Preceding undated comment added 05:53, 20 January 2016 (UTC)

• Support -- weakly, because I think the line should show sources but the discussion is saying the bigger issue is it's a bad line. Markbassett (talk) 16:59, 22 January 2016 (UTC)

Provide additional information for further context

• Could support...if done right. Flyer22 Reborn (talk) 15:44, 17 January 2016 (UTC)

• Question Have you got any specific examples of alternatives you could support? Scoundr3l (talk) 21:37, 18 January 2016 (UTC)

• Support This would no doubt be a happy compromise for all of us. I would like to at least see a range of figures. Such figures do exist, and the oft-quoted range is 68-90%, but this is from data compiled from several surveys. I am able to provide additional sources on figures if needed. Scoundr3l (talk) 21:37, 18 January 2016 (UTC)
• Support. Due to the complexity and wide divergence in methodology of the different sources, claims such as this one (which, as far as I can tell, is 100% true) should be clarified in maximum detail. --Sammy1339 (talk) 01:08, 19 January 2016 (UTC)
• Conditional Support I think the additional context is sort of provided by subsequent sections, and I think my own proposed edit is a little unwieldy because it introduces a survey-based measure of sexual assault before explaining why survey-based measures of sexual assault are necessary. Still, I think this is fine if we think it's a necessity to include more in-text information. Nblund (talk) 22:19, 19 January 2016 (UTC)
• Yes Yes YES !!!! It seems that the text is from one source and showing that is not liked so consensus to work on the wording ('numerous studies' or whatever) to something that does not draw these concerns seems more indicated than talking about whether or not to continue with what presents an appearance of a bad-WP soapboxing exaggeration. Markbassett (talk) 17:03, 22 January 2016 (UTC)

Proposed wording

The majority of rape and other sexual assault victims do not report their attacks to law enforcement. From 1995 to 2013, 67% of the sexual attacks against non-students and 80% of the sexual attacks against college students reported in the National Crime Victimization surveys were not reported to police. (citing this, already in the refs) Nblund (talk) 22:16, 19 January 2016 (UTC)

I would support that, Nblund. Thank you. Flyer22 Reborn (talk) 22:24, 19 January 2016 (UTC)

• I like the second sentence and appreciate the attempt at compromise. However, my concern is that the cited survey is only related to college-aged females in the United States, specifically 18-24. This would be approximately 15 million people in a very specific cultural demographic represented by this survey data. I'm not trying to tear down the house before it's built, but it doesn't seem like a fair representation of the initial statement. Scoundr3l (talk) 07:49, 20 January 2016 (UTC)

Here is one possibility which borrows terminology from the citation (thank you, Nblund, for providing): "Research has consistently shown that the majority of sexual assault victims do not report their attack to law enforcement."

This would then go on to cite specifics, as necessary. Notice it is not a direct attribution, so as not to immediately favor any of the 11 some-odd surveys which support the claim in the citation, but at least provides context as to where the data comes. It also plainly acknowledges the majority of the viewpoint. This is also quite open to modification, so feel free. Scoundr3l (talk) 08:48, 20 January 2016 (UTC)

I think "research has consistently shown" works, especially if subsequent statements bolster that point. Regarding the overall average: we could use the most recent NCVS available: 34% of rapes/sexual assaults reported on the 2014 NCVS were reported to police. Or we could use this report that aggregates over four years: this source "between 2006 and 2010, 65% of rapes or sexual assaults reported on the National Crime Victimization surveys were not reported to the police". I might add a couple of additional sources specific to college students after that. Nblund (talk) 18:08, 20 January 2016 (UTC)

Nblund - rewording seems the right area but what the second line shows is only that there is disagreement among reports, not a support for the first line. I'm now twigging on the topic is supposed to be campus sexual assault, so the whole first line and ensuing discussion about statistics accuracy seems a bit WP:OFFTOPIC or at least the context by title of this section "Campus sexual assault", "Prevalence and incidence of rape and other sexual assault" subsection would need to convey why is Prevalence all about how the statistics do not agree ? Cheers Markbassett (talk) 17:26, 22 January 2016 (UTC)

I'm not what you mean by disagreements. The second line in the wording I proposed says that there are differences between two different populations (college vs. non-college), but I don't think that really reflects a disagreement so much as an actual difference between the two populations. Can you clarify what you're saying here? Nblund (talk) 18:07, 22 January 2016 (UTC)

Further commentary

I think there are valid concerns being expressed by both sides here, but I also think there are a couple of important points to keep in mind.

• 1."It's not necessary" or "it's unhelpful for readers" is actually a perfectly valid reason to oppose including something in an entry. I think there are additional concerns that are more central here, but I don't think we should dismiss that consideration out of hand. The section on WP:INTEXT specifically discourages using in-text attribution when a footnote would suffice.
• 2. Not including an in-text citation is not the same as not including a citation at all, and it doesn't prevent readers from checking facts or forming their own opinions. For the vast majority of cases, a footnote citation (like the one we use) is considered sufficient.
• 3. Most importantly: survey research is science. There's no precedent for handling surveys any differently than any other form of scientific evidence on Wikipedia, and I think the argument "surveys can't be factual" would require us to rewrite vast sections of not only this entry, but of a vast swath of Wikipedia entries. Virtually every piece of demographic data cited in the entry for Demography of the United States, for instance, comes from a survey, and most of those statements do not include in-text attribution. Even if you think this should be the standard, it clearly isn't the standard right now, and imposing it would require more than an RfC. Nblund (talk) 22:52, 19 January 2016 (UTC)

I believe we have common ground on these points, but I have a few additional perspectives:

• 1. Those comments ("it's not necessary", "it's unhelpful") are indeed valid considerations. Their weight must be carefully considered, however. To an article with a stable status quo, any change at all is generally more unnecessary than necessary. Yet, in the interest of collaboration, Wikipedia encourages bold edits and discourages reversion. Only in cases where an edit clearly makes the article worse should reversion be considered, so perhaps the better question is whether or not it's necessary to exclude the change.
• 2. Indeed, there is no argument that the statement is cited. However, relating to my first bullet point above, the statement is commentary related to some sort of data. By not including the source of data in the text, it makes it all but impossible for other editors to include additional commentary on that data (which certainly exists) and thus inadvertently gives the sources sole authority over the conclusion. This does not seem to be in the best interest of the reader or the other editors.
• 3. Agreed. Sociology (forgive me if this is not the accurate term for this type of survey) is a science and it follows a method. Like all sciences, this method is generally an ongoing process of examining previous data, drawing conclusions, adjusting methodologies, and further testing hypotheses. The cited survey in particular put a great deal of effort into ensuring the accuracy and objectivity of their data. I think it's important that we give the conclusions all the respect they deserve: they are not guesses, their methods are sound, and we have every reason to believe the conclusions are accurate. However, there are still a great many debates ongoing in this area of study, notably among sociologists themselves. Even if all surveys show a greater than 50% incidence of non-reporting, it is unlikely that there is consensus on the methodologies nor is it likely that others surveys will replicate the results, making the conclusion potentially contentious. At least uncertain enough that it should be written in the commentator's voice, not Wikipedia's, or at least attribute where the data is coming from. But I'm not asking you to take my word for it, I'm only asking that we improve the wording of this statement enough that additional commentary can be easily added. If the results of the US Demographics statistics were a matter of serious contention, I have no doubt Wikipedia would accurately report on that. This subject, as it stands, is a bit more controversial. Scoundr3l (talk) 07:19, 20 January 2016 (UTC)

Wording Kentucky finding

Regarding this. I left out "attempted" but the wording should be something like: "A separate, mandatory, survey conducted by the University of Kentucky found that 5% of college undergraduates experienced completed or attempted oral, anal, or vaginal penetration while unable to consent due to drugs or alcohol, or by force".

The placement of the sentence and the emphasis on the higher response rate seem to imply that it was somehow methodologically superior to the AAU, but that's not necessarily true: low response rates don't necessarily indicate bias, and high response rates don't necessarily indicate a lack of bias. More importantly: simply saying they measured "sexual assault" is misleading. They only measure completed or attempted penetration in the last year. The 5% finding is actually higher than the most directly comparable finding from the AAU survey: (table 3-21, page 117 of the pdf) an average of 3.1% of undergraduates experienced completed or attempted penetration using force or incapacitation in the past year. As it stands, i'm not really sure why this finding is important, or why it belongs in the AAU section. It's not very well-documented beyond the brochure, and it seems sort of irrelevant to the AAU survey itself.

This edit is essentially a word-for-word reproduction of the same as the material I have objected to in the past. Reasonable people can disagree, but it's tendentious and time-wasting to keep re-introducing disputed material without making an effort to resolve the issues. Nblund (talk) 19:52, 20 January 2016 (UTC)

I think it's problematic that we lack access to the primary source and newspaper wording seems to have a bit of a willy-nilly attitude about conflating different terminology. The statistic appears to count attempted, but not completed assaults, as well as cases of intoxication which might not meet the legal definition of sexual assault. I would suggest not including it unless we can find the original study, or at least making sure our wording closely matches what's written in the brief summary that U Kentucky made public. --Sammy1339 (talk) 21:20, 20 January 2016 (UTC)

I tend to agree, but that approach may receive some pushback from others. The full report was supposed to come out in December, but it still hasn't been released as far as I can tell. Based on their public statements, it sounds like they used questions similar to the ones used to measure rape on the Campus Sexual Assault study. If that's the case, their questions about incapacitated rape would conform to the legal criteria for most states. Still, I think you're right that, without access to more information, there's a risk of mis-characterizing things. Nblund (talk) 16:29, 21 January 2016 (UTC)

Here is what I am reading in the linked brochure: "[4.9% of UK students] reported unwanted sexual experiences (vaginal, oral, or anal sex) that occurred because: they were slipped drugs or alcohol and unable to consent; they were unable to consent due to voluntary drinking or taking drugs; they were threatened with physical harm; or they were physically forced. This percentage also includes individuals for whom someone attempted to force them to have sex, but they were able to escape." I agree with more specific wording, as opposed to "any sexual assault", as the more details provided the better the context for the reader. However, I would even suggest we change it to "unwanted sexual experience" so as to match the text and so there are no squabbles over the definition of sexual assault. Also, it doesn't seem necessary to round this number to 5% if I'm reading the relevant content, as often these numbers have already been rounded from a more specific decimal. The second sentence does indeed seem lax in its wording, but I think it's enough that we mention that this includes attempted. Given all that, my personal wording for this sentence would be "...encompassing 80% of students (24,300 respondents) found that 4.9% of that school's students in the past year were victims of unwanted sexual experiences, including completed or attempted oral, anal, or vaginal sex without their consent." I hope that helps. Scoundr3l (talk) 22:34, 20 January 2016 (UTC)

I just realized that "including" may be contentious phrasing, but I have no objection to "defined as". Scoundr3l (talk) 22:37, 20 January 2016 (UTC)

No, this is exactly the point. It should state clearly exactly what is being measured, and this is required by WP:PRIMARY. Your abbreviated wording, omitting e.g. "unable to consent due to intoxication", is original research. --Sammy1339 (talk) 22:59, 20 January 2016 (UTC)

I'm not sure I understand where your objection lies. My statement mentioned exactly what the text said was measured and mentions that it was without their consent. The various methods that may produce a lack of consent seem unnecessary to that point, but you're welcome to include intoxication if you like. Surely some amount of content can be reasonably omitted for the sake of WP:PARAPHRASE, so I'm not sure what you qualified as OR, but otherwise we might as well make it a direct quote. Scoundr3l (talk) 00:22, 21 January 2016 (UTC)

Because differences in operational definitions and standards of consent are key to this topic, it's critical to completely describe them. Otherwise meaning is lost. This would be a very bad use of WP:PARAPHRASE (which is an essay, not policy). The situation here is closer to the sort of thing WP:MEDRS was written to deal with. In medicine, epidemiology is a notoriously fickle subject, which is why MEDRS warns against using primary sources at all if it can be avoided. Although sexual assault poses serious dangers to physical and mental health, this is arguably not a topic where MEDRS applies; nevertheless, we are dealing with similar sourcing issues. --Sammy1339 (talk) 00:31, 21 January 2016 (UTC)

Are you objecting to the source or only to the paraphrasing of the source? If the latter, what else would you include that would be an accurate representation of the source, in your opinion? Scoundr3l (talk) 00:55, 21 January 2016 (UTC)

If we are including this survey, I would prefer to be explicit about what was measured. My concern is that it looks like we're making a comparison between Kentucky's findings and the findings of the AAU, even though those results really aren't comparable because they measure different types of sexual violence. Nblund (talk) 16:22, 21 January 2016 (UTC)

It should be a two way street then. Nblund wants to be explicit about the Kentucky Survey, but wants to minimize the detailed findings of other surveys that tout higher numbers, including how often the "victims" don't think it's serious enough. Here's the theme - report high level numbers (per the lede), with no attribution about what and how they are measuring it. Also, try to argue that surveys that have very high rates are just as good, no, better than surveys that indicate lower rates. See the theme?

Regarding Nblund's queries about the differences in methodology between the AAU and Kentucky studies, the AAU study used responses from non-random and voluntary sample with a low participation rate. This creates the risk of Sampling bias. Per the AAU study itself, "An analysis of the possibility the estimates were affected by non-response bias found that certain types of estimates may be too high because non-victims may have been less likely to participate." The Kentucky survey represents nearly the entire population, was not voluntary, which means it's more representative of the student body.Mattnad (talk) 13:54, 21 January 2016 (UTC)

Yeah, this is exactly what I think we should avoid trying to imply. A sample with a high response rate can potentially be less representative than a sample with a low response rate and vice versa. Whether or not a response bias exists is actually a pretty tricky empirical question: notice that the AAU report dedicates about 30 pages (appendix 4) to analyzing the sample for bias. The KU survey may be worse, or better, but they either didn't check for bias, or haven't released those results. It's really speculative to say that the survey is of better quality. It certainly isn't something that you could cite a reliable source to support. Nblund (talk) 16:22, 21 January 2016 (UTC)

It's more than an implication. It's a basic fact about surveys and sampling and it's overtly stated by the AAU study designers themselves. I'm confident you know the difference, but let me explain to other editors. The closer the sample is to capturing the entire population the better. If you can poll everyone, then the results reflect the entire population. If you cannot capture the entire population, you then need to control for sample bias. One way this is done to is make sure the sample is random. Kentucky does a good job because it captured more than 90% of the population, and did not have any issue with self-selection bias because it was mandatory. AAU is closer to 15%, and was not random. So 85% of students did not respond to the survey, and of the 15% who chose to respond, the AAU itself states they may be more inclined because they had been assaulted. I don't think you really believe the two approaches are equivalent, but you want readers to think they are. You carefully avoid the Complete Sample, vs. Random Sample vs. Self-Selected Sample issues to make your argument, and that's just misleading. Mattnad (talk)

Yeah, it's not true that higher response rates are always better. There's actually quite a bit of recent research on this topic that suggests that bias can be the same or even higher in surveys with very high response rates. The AAU authors cite some of this literature in the report (footnote 1 on page vi of the introduction). The AAU does say that there is evidence of a response bias on their survey, and that is noted in the entry. However there are no indications of the quality of Kentucky's data, and the claim that one is superior to the other doesn't appear to be something that you can support with a reliable source. Nblund (talk) 17:07, 21 January 2016 (UTC)

Here's a very specific source commenting on the KU survey which I quote:

"Until now, the school has had to rely on information that was volunteered. Follingstad explained that volunteered data is not always representative. “A survey that goes out to a campus is relying on whomever is willing to complete it,” she said. “There is always a concern that samples are skewed.” That’s why campus officials decided the survey should go out to entire student body."

Can we drop this now? AAU study was voluntary. Kentucky was not.Mattnad (talk) 17:58, 21 January 2016 (UTC)

Where does it say that the data quality is superior to the findings from the AAU? Where does it say that they have empirical evidence of reduced response bias? That statement explains Kentucky's reasoning, but it doesn't say anything directly about the quality of the current data, and it certainly doesn't offer any support for making a comparison across data sources. This is a very complex statistical question, it's not really something that we can reasonably speculate about. Nblund (talk) 18:32, 21 January 2016 (UTC)

Now you're just being difficult. The AAU designers say their study is skewed. The KU say theirs is not, with the same reasoning.Mattnad (talk) 18:57, 21 January 2016 (UTC)

I'm not trying to be difficult, but it seems like you're putting words in the mouths of these researchers. The AAU says they have some indications of a response bias. They draw this conclusion after running some fairly rigorous empirical tests. KU says that volunteered data isn't always representative, but they don't actually assert that their results are free of bias, and they absolutely don't make a comparison between data sources. Doing so would be pretty irresponsible, because they don't appear to have analyzed the data for response bias -- at least not yet. Again, this is a question that requires some fairly sophisticated statistical know-how and technical skills, it would be pretty clear-cut OR to assert something like this without a reliable source. Nblund (talk) 19:35, 21 January 2016 (UTC)

Again quoting, "There is always a concern that samples are skewed.” That’s why campus officials decided the survey should go out to entire student body."" What part of that do you not understand? How about I add what KU and the AAU people say about their own studies then, with quotes? Or will you object to that?Mattnad (talk) 20:06, 21 January 2016 (UTC)

It would depend on the wording and arrangement, if you arrange the quotes in a way that appears to be making a comparison between the two, I would say that still poses the same problem. Others are questioning whether this belongs here at all, so maybe you should propose something above and see what others think.

The AAU is making a statement based on a fairly rigorous response analysis. Follingstad, on the other hand, is discussing how KU's novel approach might alleviate the issue, but it's not something that has empirical support, nor does it suggest anything about the AAU. I think this comparison is also a stretch, but KU does compare their results to the Campus Sexual Assault study. Their findings actually indicate a higher rate of sexual violence than the CSA (5% in the last year vs. 3.4% since entering college). Nblund (talk) 20:51, 21 January 2016 (UTC)

The AAU study is clearly flawed due to the low response rate, while we just don't know enough about the U Kentucky study. There are problems with mandatory surveys too, such that some students may not read the questions and just try to get it over with. I don't think we can reasonably compare the results unless we know that the methodology was similar, and I'm bothered that we don't seem to have access to the original study, just the summary of results. --Sammy1339 (talk) 21:52, 21 January 2016 (UTC)

Yeah, there's also the problem that the 80% completion rate appears to include students who completed the survey but selected "prefer not to answer" for certain questions. The number of valid responses for each item is probably lower than the total participation rate. Again, this is something that would probably be laid out in the full report.

Looking over the entry, I think the section on "prevention efforts" probably should contain some information on the strategy of using climate surveys, and I do think some info on the Kentucky, Michigan, and Yale approaches (as well as the AAU) might be useful there. Kentucky did get some press coverage for the idea of making this survey mandatory, and I think that approach is noteworthy even if we're unsure about how to deal with the results themselves. Nblund (talk) 18:34, 22 January 2016 (UTC)