Talk:The Keys to the White House

Pre-1980 elections

@Trajan1: I see you attempted to list all the retroactive predictions that Allan Lichtman and Keilis-Borok made to develop their prediction model. I deleted it because it was rather messy. If it interests you, I started a similar table in my Sandbox: https://en.wikipedia.org/wiki/User:Kurzon/sandbox#Keys_to_the_White_House. If you really want to go through with this, I suggest you start with my table as a template since you seem to struggle with making tables.

I will point out that Lichtman didn't really predict the pre-1980 elections, rather he used the data from those elections to develop his prediction model. Once he had a model that retroactively predicted past elections, he used it to predict future ones. Kurzon (talk) 20:31, 20 June 2022 (UTC)

yeah, sorry. I'm just an amateur editor who's good at visual editing. I did use the primary source from Lichtman's book for the retroactive model, if you were curious, so it was not unfounded. I should've done a better job at creating the table; that's my bad. I suck at coding.... :( Trajan1 (talk) 13:49, 23 July 2022 (UTC)

No need to be so apologetic, you didn't break anything (permanently). I thought of including the retrospective table as part of an in-depth explanation into the mathematics behind the Keys. In the end, I decided it was not really useful. Are you better at maths than you are at coding? Perhaps you can read Lichtman's original paper and make sense of it. (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC349231/) Kurzon (talk) 12:20, 24 July 2022 (UTC)

I'm not good at coding, but I thought that for what it's worth, the table was really useful and looked fine. BazingaFountain42 (talk) 21:27, 12 July 2024 (UTC)

The strong short term economy key

Lichtman said the strong short term economy key was likely true, not certainly true. But this article marked this as true anyway. Can someone fix that? 100.16.156.64 (talk) 17:46, 7 June 2024 (UTC)

The day before you made this comment, Litchman said that it is certainly true

https://www.youtube.com/live/OGjuweWtu9Q?t=971 Botaeditor (talk) 00:44, 30 June 2024 (UTC)

My mistake. He did mark it down as certainly true. While I disagree with his assessment about the likelihood of a recession, what matters for this article is what he says, not my opinion 100.16.156.64 (talk) 23:14, 9 July 2024 (UTC)

2016 Election Confusion

It seems that there is some confusion about the 2016 prediction. Though the keys and Lichtman predicted a Trump/Republican win in 2016, this was about the popular vote, which Trump/Republicans did not win.

There are some changes that should take place:

1. A clearer indication in the page regarding what happened.
2. Some indication within the table/prediction record.
3. Addition to the "Criticism" subsection.

I will look for some sources. Caraturane (talk) 14:48, 23 June 2024 (UTC)

There are several sources regarding this discrepancy. Is anyone able to verify according to this source, a Social Education paper in Volume 80? This appears to be the source which is most frequently cited. This is also explained here, and here. A citation to his 2016 novel would also be consistent with this inconsistency, there is a quote in it: “predict only the national popular vote and not the vote within individual states.” Caraturane (talk) 18:28, 23 June 2024 (UTC)

When Lichtman made his prediction for 2016, he simply said Trump would win, he didn't talk about the popular vote. So I mark that as a qualified hit. And to show I am fair, I marked his 2000 prediction as a qualified miss. He did not discuss the popular vote in the journal articles wherein he made his prediction, he simply predicted Gore would win. He also did not and could not predict the miscounting of votes in Florida. Kurzon (talk) 22:14, 23 June 2024 (UTC)

His books from the 1990s do say popular vote only, so I am inclined to count 2000 but not count 2016, consistent with the sources. Are you able to get access to this source? It is from October of 2016 and has been brought up as saying "popular vote" but I cannot get access. I may try emailing him myself to see if he can send it over, it seems like this source specifically would answer a lot of the confusion. Caraturane (talk) 22:29, 23 June 2024 (UTC)

That's what his books say, but his books don't actually contain predictions, they explain the theory behind the predictions. Even if there is a discrepancy between his theory and his prediction for 2016, he nonetheless made a correct prediction. Kurzon (talk) 08:34, 24 June 2024 (UTC)

The October 2016 paper allegedly does, but I cannot find a copy. Do you have any ideas? Caraturane (talk) 21:34, 24 June 2024 (UTC)

I am hoping to read a copy of that as well. In the meanwhile, the "The Postrider" article you linked to does contain a quote from that article:

"the Keys predict the popular vote, not the state-by-state tally of Electoral College votes.” Apprentice57 (talk) 07:13, 25 June 2024 (UTC)

https://www.dropbox.com/scl/fi/ksm9n3qsptqkjd7avr5km/Allan-Lichtman-2016-The-Keys-to-The-White-House-Social-Education-80-5.pdf?rlkey=qdtcni8kxv60qxfz52ewecy1u&st=27o5ep7i&dl=0 Kurzon (talk) 13:57, 25 June 2024 (UTC)

Wow good find, Kurzon and thank you for your help. We should rely on this citation concerning his 2016 prediction. It is his last and most definitive published word on the subject I was able to find (later than the Washington Post article, which is less clear than the language used here). This part makes it very clear that this was his prediction, when combined with the statement two paragraphs before that this only applies to the popular vote: "Since that time, however, the Keys have shifted and now point very slightly to a Republican victory in 2016... Although the model narrowly predicts victory by a generic Republican candidate, Trump is anything but generic and may vitiate that prediction." The first part of the article also "The model has successfully predicted the results of the popular vote in all eight American presidential elections from 1984 to 2012." Tomcleontis (talk) 17:14, 25 June 2024 (UTC)

A thank you from me for sharing that as well. After reading it I feel that the quote that "The Postrider" selected from it that I mentioned above is fairly smoking-gun and not mollified by context. Curious as per your thoughts Kurzon. Apprentice57 (talk) 19:42, 25 June 2024 (UTC)

Lichtman missed one prediction and it was either 2000 or 2016 depending on how you look at it. Most people know Lichtman's prediction from his TV appearances. Who reads obscure science journals? I bet most academics didn'tread that article in Social Education. Lichtman went on TV and told everyone Trump would win. I am inclined to go with that.

Oh, and I am down $20 for the article. No big deal, I would have wasted that money on chocolates otherwise. But that I couldn't find it on Libgen goes to show how obscure the paper is. Kurzon (talk) 21:00, 25 June 2024 (UTC)

THANK YOU KURZON! That's very helpful. I have a few thoughts:

1. This paper is much clearer than I expected. I did not expect it to explain it was the popular vote several times, that's pretty open and shut.
2. This paper is his last word available on the subject and it is in line with his official predictions each year, this is where he publishes his prediction and its updates, thank you @Tomcleontis for looking at that.
3. Regarding his own words. We should value a published paper by him more definitively than a TV interview, and we should value something from October more than something from September, is that really up for debate?
4. I have looked at this for a while now and this reminds me of the dispute over when he changed his predictions. He said for a while that it was after 2000, but sometimes says something else, he is an inconsistent source on this, and a biased one as well. I want to be clear that this is nothing against him personally but when it comes to materials which review his work, they should be independent rather than by him. For an academic, peer or independent review.

Caraturane (talk) 21:35, 25 June 2024 (UTC)

Consistent with best faith and independent review, we could just rewrite much of the article to explain that he's believes one thing, but an independent review has shown another thing. We can use his own papers and writings to make this point, and use the Postrider article to tell the complete picture. That is the most honest thing to do, but I think at the end of the day you are right, we should weigh the article, his October 2016 paper, and the word of others more than his own word. Tomcleontis (talk) 13:04, 26 June 2024 (UTC)

The books do contain his predictions. The only caveat is that for 2016, the book was published in May so a couple of factors were still uncertain at the time of publication. But the fact he's going through the motions at all show that he didn't change the keys between the book's publication and when he would later in media/a paper fill out the uncertain factors.

Conveniently for Lichtman, the 2020 book didn't contain the full 2016 prediction without uncertainties owing to an error. Which would've been useful clarification here. The linked "The postrider" article above goes into this a bit more. Apprentice57 (talk) 06:37, 25 June 2024 (UTC)

I think of the books as persuasive evidence of his predictions and how they work, but the Social Education paper is his actual predictions. Tomcleontis (talk) 17:14, 25 June 2024 (UTC)

Edit to add: the 2016 book raises eyebrows but this paper Kurzon found above (and mentioned in the Postrider article) confirms. The page should cite to all three: book, article, and paper. Tomcleontis (talk) 17:16, 25 June 2024 (UTC)

Hey there, author of the reddit post linked to by Caraturane here. I do have access to the (May) 2016 book in question in digital form. I do not think it is freely available, but if it is of help for you/them/others the ISBN is 1442269200.

I'd like to quote a few sections from said 2016 book (italics are mine, bolding is theirs). I apologize in advance for the overall length but I want to give as much context as possible. Unfortunately the page numbers are inconsistent with the print book as I am using a digital copy:

1. "The keys to the White House focus on national concerns such as economic performance, policy initiatives, social unrest, presidential scandal, and successes and failures in foreign affairs. Thus, they predict only the national popular vote and not the vote within individual states. Indeed, no system could have predicted the 537 vote margin for George W. Bush in Florida that decided the 2000 election." - Page 11
2. "Each of the thirteen keys (see page 3) asks a question that can be answered yes or no before an upcoming election. To avoid the confusion of double negatives, the keys are stated as threshold conditions that favor reelection of the incumbent party. When five or fewer keys are false, the incumbent party wins the popular vote; when six or more are false, the challenging party prevails." - Page 14
3. "Because the keys to the White House diagnose the national political environment, they correlate with the popular balloting, not with the votes of individual states." - Page 14
4. "Based on the historical odds since 1860, the chances are better than twelve to one that the popular and electoral college vote will converge in any given election. However, these odds presume continuity over time in the relationship between popular and electoral college votes. Some analysts have suggested, however, that this relationship may have changed given the sharp division in America between Republican “red states” and Democratic “blue states.” - Page 11
5. "Only three times since 1860, however, has the electoral college not ratified the popular vote: the “stolen” election of 1876, when Democrat Samuel J. Tilden outpolled Republican Rutherford B. Hayes 51 to 48 percent but lost a disputed contest for the electoral vote; the election of 1888, when electoral college votes overrode President Grover Cleveland’s narrow popular-vote margin over Benjamin Harrison; and the 2000 election described above." - Page 14
6. "THIRTEEN KEYS TO THE WHITE HOUSE The keys to the White House are stated as conditions that favor reelection of the incumbent party. When five or fewer statements are false, the incumbent party wins. When six or more are false, the incumbent party loses." - Pages 15-16

My read from these is that 1, 2, and 3 establish that the keys predict the NPV. 4, and 5 make the case that the NPV is very strongly correlated to the EV and thus overall winner. That is why Lichtman then introduces the Keys themselves, he uses the abbreviated language "wins" without literal contradiction with the text.

I think that may explain why there's confusion here. If you just look at the snippet that only introduces the keys (page 15-16) then the language implies just EV. But the extra context, in particular quote 2 above, clarifies that Lichtman means popular vote, but believes the two nearly if not totally equivalent.

The books also do contain predictions about the upcoming election cycle (the 2016 election is on pages 230 - 235 for instance), but with some uncertainties owing to being published too early in the election season. I don't believe there's a reasonable narrow exception that Lichtman could've changed the system to EV between the May 2016 book publication, and when he finalized the uncertain keys a couple months later.

Also see quote 1: ignoring Lichtman's claims and just analyzing the keys in and of themselves, it makes plenty of sense that national factors can only predict a national statistic like the NPV (rather than one that changes based on state specific ones like the EV).

I am therefore of the strong opinion that we should mark the keys correct for 2000, but incorrect for 2016. They always have, and always will predict popular vote. Apprentice57 (talk) 06:33, 25 June 2024 (UTC)

THanks for all that research, Caraturane and Apprentice57. Hopping in to add: the problem here is that to assume he predicted 2016 right, we'd have to believe that he changed the system and what it means to win without telling anyone before the 2016 election (which we cannot, because we should only be putting substantiated facts on the Wikipedia page). I do believe he changed his predictions at least after 2016 though. Therefore, I think we should call 2000 "questionable" (his 2000 prediction wasn't explicit enough for my liking, but I get the argument for including it) and 2016 wrong (the October 2016 paper and the independent reporting pretty much proves this). Tomcleontis (talk) 13:08, 25 June 2024 (UTC)

I wouldn't oppose a questionable rating for 2000.

Going forward after 2016 he is on record about them predicting the EC vote I concur (it doesn't really make sense *why* that's the case given he hasn't changed the wording at all, which is why I say they always will predict the popular vote, but that's at least a methodological issue and can be tackled later). Apprentice57 (talk) 18:18, 25 June 2024 (UTC)

I agree. The keys were correct in 2000 and incorrect for 2016. Caraturane (talk) 21:37, 25 June 2024 (UTC)

In 2000, the Keys were correct, Lichtman was not. Kurzon (talk) 21:50, 25 June 2024 (UTC)

His actual line in the September 2016 interview he relies on is: "Based on the 13 keys, it would predict a Donald Trump victory." I think it's fair to conflate the Keys prediction with his own. Caraturane (talk) 21:57, 25 June 2024 (UTC)

He is one with the keys!

No but seriously this is a good point.

Tomcleontis (talk) 13:05, 26 June 2024 (UTC)

Hi all, I've gone ahead and begun making these changes, relying on the paper, book, article, and Twitter post from Nate Silver. I explained that Lichtman feels he did predict 2016 as well. This probably requires much more work on the table in particular. Tomcleontis (talk) 13:38, 27 June 2024 (UTC)

I am of the position that if you put a note that his keys were "wrong" in 2016 because they predicted the popular vote, you should also put the note for 2000 saying he was actually right. Let's just be consistent here. Wikiman5676 (talk) 16:47, 28 June 2024 (UTC)

You are right, I added that on the table. @Tomcleontis raised a point that the table could probably use some work generally given this dispcrepancy and the fact that he changed what was being predicted after 2016, so I'm open to some ideas here too. Caraturane (talk) 16:55, 28 June 2024 (UTC)

I agree for a larger revision of the table. I think it's important that we show using the existing red/blue cell coloring that the prediction matches the outcome in 2000 (both blue for Gore) but didn't match in 2016 (red for Trump popular vote prediction on the left and then blue for a Clinton popular vote victory on the right).

I'm not sure how to best show that what was predicted changed starting in 2020. If the table wasn't so wide already I'd suggest adding additional columns.

Perhaps we could end the current table at 2016, and also have the last two columns instead say "Predicted Popular Vote Winner" and "Actual Popular vote Winner". Then add a second table below that has entries for just 2020 (and 2024) but with the last two columns saying "Predicted Winner" and "Actual Winner". Still clunky but perhaps better than an even wider single table. Curious as to others' thoughts. Apprentice57 (talk) 19:10, 28 June 2024 (UTC)

I like the idea of having columns that mention both. I'm also okay with how the table is now with the actual winners noted and the popular vote winners put in as a note. Wikiman5676 (talk) 19:52, 28 June 2024 (UTC)

I prefer to keep the table as it is, and use the paragraphs above to explain the nuances of the 2016 and 2000 elections. Kurzon (talk) 12:00, 29 June 2024 (UTC)

I don't think that's a good option, it's not really accurate now as is (or well, is with the note but that's clunky). Apprentice57 (talk) 13:06, 29 June 2024 (UTC)

An idea might be to color 2016 and 2000 differently and leave an asterisk, to show that they are different? Also an asterisk at the top of the table in the "Predicted Winner" and "Actual Winner" headers which reads: "Prior to 2020, the Keys predicted the winner of the popular vote." would help clarify. Tomcleontis (talk) 13:12, 1 July 2024 (UTC)

Forgive me if I was presumptuous, but I figured that with a little change of language we could efficiently describe the anomalies in Lichtman's record with just a couple of sentences in the lede. Please see my edits and tell me what you think. Kurzon (talk) 17:41, 11 July 2024 (UTC)

For the time being I think we should lean towards more thoroughness and more sourcing about the discrepancy, as that seems to be the source of all the dispute. I also think saying "explained below in the article" is inappropriate language for a Wikipedia article and I think as it was (with sources provided) was a good compromise. I do think maybe including a section on this in the "Criticism" section may be helpful as well, though I realize that is adding more. Tomcleontis (talk) 17:53, 11 July 2024 (UTC)

The way it was written (before) was a bit confusing, so I cleaned that up too. As he did switch what he was predicting after 2016 anyway, so we should not say "9 out of 10" when referring to the popular vote (as it was 8 out of 9, then 2020). It may still need some cleanup. Tomcleontis (talk) 18:13, 11 July 2024 (UTC)

Regarding Lichtman's Prediction Record

I wanted to ask if it would be possible for me to restore the deleted portions of the article regarding Lichtman's prediction record?

I think that removing almost all the details on his predictions outside of 2000 and 2016 leaves out some pretty important context regarding his prediction history. E.g. That his 1984 prediction was was made two years before the election or that his 1988 prediction was made when Dukakis led Bush by 17%.

As is, it doesn't even list when he made his first prediction.

I just want to see if I can restore the information, possibly in trimmed down form.

I found it very helpful personally, especially the articles cited. DanielXW1 (talk) 17:51, 5 July 2024 (UTC)

I support this, would be cool to have a graph or something too on how early some predictions were made as the 1988 one does stand out. Tomcleontis (talk) 20:09, 9 July 2024 (UTC)

I don't like it, but I'll leave it to a vote. Kurzon (talk) 21:38, 9 July 2024 (UTC)

I agree. Saying when he made the prediction is a good/relavent column to put. Wikiman5676 (talk) 17:25, 15 July 2024 (UTC)

RFK Jr's polling aggregate average

The note that says "As of July, Lichtman's polling aggregate average is 8.8%" seems irrelevant, because Lichtman cares more about whether polls are consistently above 10% than about polling averages. 100.16.156.64 (talk) 23:17, 9 July 2024 (UTC)

It is relevant because it shows how far off they are. Pollwatcher1234 (talk) 15:57, 17 July 2024 (UTC)

Right, but I think “they are” really means “tend to be” here. Not that it stomps on the point at all. RobertGustafson (talk) 01:52, 3 August 2024 (UTC)

Has anyone applied the keys for pre-1860 elections (esp. 1856)?

I notice that the keys were not applied retrospectively to elections before 1860–including the 1856 election, the first featuring Democrats versus Republicans. Is this because the model breaks down pre-1860—say, because of a lack of reliable data? Or has no one actually tried to test it for earlier races? If anyone HAS done the analysis, then we should consider incorporating it. 24.154.117.91 (talk) 02:29, 21 July 2024 (UTC)

Lichtman said that he started in 1860 because before that there was different parties and the dynamics were different. Anything before 1860 would not be part of Lichtman's system and thus irrelevant to this article regardless of election results. RevDan (talk) 09:10, 27 July 2024 (UTC)

1856 was the first election where the 2 major parties were Democrats and Republicans. The parties were NOT different then, but they were prior to 1856.RobertGustafson (talk) 23:56, 2 August 2024 (UTC)

As to the “dynamics”, how does Lichtman say they were “different” prior to 1860? (The parties were only different prior to 1856.) And I still ask, has anyone tried to apply the keys to previous elections, or has Lichtman said that the results of doing so were unreliable (which would be a good reason for not starting earlier)? RobertGustafson (talk) 00:03, 3 August 2024 (UTC)

Lichtman keys wrong

Bottom row wrong for the keys. There were scandals, trump is arguably charismatic, there were major scandals for incumbent, economy is in recession etc. needs update — Preceding unsigned comment added by 172.116.60.189 (talk) 22:20, 22 July 2024 (UTC)

Unemployment is below 5%, and Lichtman defines “charisma” as being inspirational and having large appeal beyond one’s party base—which Trump doesn’t go have. 24.154.117.91 (talk) 07:45, 23 July 2024 (UTC)

“Go have” was supposed to simply be “have”. 24.154.117.91 (talk) 07:46, 23 July 2024 (UTC)

BTW, I personally would note after hearing Kamala Harris and the way many people describe her style of stump-speeching that SHE could be described as “charismatic”, but it remains to be seen if her way of speaking and presenting a case has appeal beyond HER party base—let alone is described so by Lichtman. Since it’s HIS model, his opinion matters most as to any candidate’s supposed “charisma”. 24.154.117.91 (talk) 20:08, 23 July 2024 (UTC)

I would argue that ONLY his opinion matters for this article. AS this is about his Keys and the predictions he makes NOT about what other people do with the Keys RevDan (talk) 09:13, 27 July 2024 (UTC)

Right on. When I said “matters most”, I meant that we should defer to Lichtman’s judgement, so that might as well be “matters ONLY”.RobertGustafson (talk) 00:05, 3 August 2024 (UTC)

There have not been any found instances of corruption, Trump charisma was not marked in previous years, and the economy is unequivocally not in recession Wigglebot23 (talk) 21:25, 23 July 2024 (UTC)

2024 key standards before official prediction.

I am seeing a lot of back and forth in the editing of the 2024 keys pre-prediction. Since this is Lichtman’s personal method of prediction, only he can have authority on what constitutes a true or false key. Can we mandate citations for any lean/likely keys, preferably based on Lichtman’s statements? Can we also standardize the false key range? Right now, I notice a range such as “2-5 false keys” taking into account both definitely false keys and lean/likely false keys, but this is misleading as lean/likely true keys could turn false, so “2-7” would be more accurate and greater show the unpredictability of the race at this point pre-prediction. Hypercorrection100 (talk) 22:45, 24 July 2024 (UTC)

I agree with you about the number of false keys. @NekoKatsun and an IP user have reverted this, so they may want to explain their opinion here? TWM03 (talk) 20:26, 25 July 2024 (UTC)

Thanks for the ping. Personally I was just counting false and leans/likely false keys. If it doesn't lean false, counting it as possibly false feels confusing, although I'm happy to defer to others on this. At this point, perhaps it's better to not include a range at all since not all keys are set? NekoKatsun (nyaa) 20:35, 25 July 2024 (UTC)

@NekoKatsun I would not object to removing the total altogether, but getting that to stick might be difficult. If any range is listed then I think 2-7 is the only one that makes sense, as if the range were actually 2-5 then the election would already have been called for Harris. Also, I think it is more confusing to have the lower end of the range as 2 (suggesting that all 3 of the leans false keys could change to true) while not similarly accepting that both of the leans true keys could change to false and putting the higher end of the range as 7. TWM03 (talk) 20:52, 25 July 2024 (UTC)

I have it as “3-7 (leans 5)” under false keys, and “TBD (leans Kamala Harris)” in the winner column—in order to show the most precise assessment from Lichtman. 24.154.117.91 (talk) 00:25, 27 July 2024 (UTC)

The 1887–1888 recession

What about the 1887–1888 recession that we have listed over on List of U.S. recessions#Free Banking Era to the Great Depression (1836–1929)? By the key's criteria of a recession during or right before election year, that should have turned the "short-term economy" key negative in 1888, thus giving Grover Cleveland six keys and predicting his defeat. All Hallow's Wraith (talk) 23:54, 24 July 2024 (UTC)

There are a few other recessions in the two years before a number of elections that he didn't list off under short-term economy, either. Those added keys would not have changed the election result, however. Unlike with Cleveland in 1888, where they would have correctly predicted his defeat. All Hallow's Wraith (talk) 09:18, 26 July 2024 (UTC)

From section: "The [economic] upturn lasted until early 1887, to be followed by a brief and relatively mild contraction that ended in the early spring of 1888. The election campaign was waged during an expanding economy, gaining for the Cleveland administration both economic keys." If a recession ended and there were two quarters of growth it wouldn't have turned key. --Shivertimbers433 (talk) 18:21, 26 July 2024 (UTC)

Well, perhaps that would be a consistent clarification to the short-term economy key. A recession in the last two years of the term. Like I said, using Wikipedia's list of recessions, it wouldn't change the outcome of any of these elections, except to correctly predict 1888 going for Harrison. All Hallow's Wraith (talk) 08:13, 27 July 2024 (UTC)

2024 data in table should be as precise as possible, per Allan Lichtman’s assessments

I think it’s important to note that while Lichtman has yet to make a final determination as to the 2024 race, he says it “leans Kamala Harris”; also, it should be noted that while the range of false keys is 3-7, the current lean is 5. I will keep putting this data in the table unless Lichtman changes his assessment. 24.154.117.91 (talk) 00:21, 27 July 2024 (UTC)

As I wrote in my edit. It was 2-4 for months. Biden dropped it became 3-5. You can't just change it to 3-7 (Leans 5) without consulting, and THEN put a "don't change unless you consult talk page" to lock it. That's not how that should work. Also it IS 3-5 false, 5-8 true. That is the logical way to count it. It's not "not-true" it's "false". It's not "not-false" it's "true". The reasonable way to count it is by counting straight false or true, as has been counted and upheld for months. Not by counting the "not-true" equates to "false", that's not the same calculation. Even Lichtman in the recent video today said that it strongly leans Harris unless something goes incredibly wrong. That dictate a 3-5 lean, NOT a 3-7 lean. But that's besides the point, because the point is how it has been counted and how it should be unless we have a discussion. You can't change it and THEN lock and then require discussion, that's not the process. Here's Litchman's pre-convention prediction posted today: https://youtu. be/zcxR-bct_Nk?si=VenqLE86T1_gaI-s&t=878 Zombie Philosopher (talk) 23:54, 29 July 2024 (UTC)

Hi @Zombie Philosopher. The page is not locked. You can see a section called "2024 key standards before official prediction." above where this has been discussed, and until your comment there was an agreement that it should be 3-7. I put the note there to prevent edit warring, as most of the editors who were changing it were unregistered users who cannot be pinged in the talk section.

In the context of this model, "not-true" and "false" are the same thing, in that by the time the prediction is made it is not possible for any key to be anything other than true or false. "3-5 false keys" is misleading in that it suggests that there is no way Trump could be the predicted winner, and "5-8 true keys" is also misleading in that it suggests that Harris would only be the predicted winner in the best possible case for the Democrats. If we had two columns, one for # true keys and one for # false keys, then your suggestion would make sense, but there is no need for that because you can work out the number of true keys simply by subtracting the number of false keys from 13.

I believe that the version with "3-7 (leans 5)" and a note to dissuade people from edit warring over it should be restored. TWM03 (talk) 08:54, 30 July 2024 (UTC)

Another option, which fits better with what this secondary source (https://www.msn.com/en-gb/news/world/i-m-a-us-political-scientist-here-s-who-i-think-will-win-the-election/ar-BB1qSSOA) is saying, would just be to have "leans 5" as the total without a range. TWM03 (talk) 12:54, 30 July 2024 (UTC)

Primary sourced, original research

Can somone explain to me why we spend so much time in this article with lengthy republication of primary-sourced original research by the creator of these keys? An encyclopedia article doesn't need to reexplain the entire theory and cases detailed in his publications. We'd be much better of focusing on independently-sourced discussion (and critique) of these keys, not just tables that simply rely on his opinions on these matters. Seems awfully crufty to me. --ZimZalaBim ^talk 21:42, 1 August 2024 (UTC)

It’s his model, so his opinion matters more than anyone else’s as he’s the one doing the predictions every 4 years! RobertGustafson (talk) 00:13, 3 August 2024 (UTC)

Also, if anyone has critiqued his keys or assessment thereof in any election, then that’s pertinent—provided it is not conflated with Lichtman’s own, or given disproportionate weight—as the article’s principal focus his his model and its track record. In particular to the tables, the views of others should either be separate from them or in citations within—as the tables’ point are to give his assessments and track record; we should defer to his views there, all the while noting any “reliable” agreements or disagreements in a way that’s visible but not overshadowing. RobertGustafson (talk) 01:47, 3 August 2024 (UTC)

So, where do these lengthy paragraphs describing the 13 keys with examples that are largely unsourced coming from? His publications? Then cite it. Otherwise these smell like unsourced original research to try to fit his model. We don't need to be making his arguements for him, just reporting on the fact these keys exist based on his formulation. Further, most of the blocks in these tables are unsourced. Where are these determinations being made and by whom? --ZimZalaBim ^talk 02:16, 3 August 2024 (UTC)

Why is “foreign policy success” for 2024 “TBD” and why are there 2 TBD columns for the winner?

Previously foreign policy success for 2024 was “leans false” with an Allen Lichtman citation; now it’s TBD with no citation. Also, splitting the Projected/Actual Winner column into 2 columns both being TBD—one with “(leans Kamala Harris)”—is redundant; either there should be 1 column with both TBD and the parenthesized lean, or there should be 2 columns, the 1st being the lean and the 2nd being TBD! 24.154.117.91 (talk) 04:45, 2 August 2024 (UTC)

The first column is TBD (likely Harris) since Lichtman has not made his final prediction yet but sources state he is likely to predict Harris to win. The second column is TBD with no lean because the election hasn't happened yet. There are two columns because only putting one column suggests that Lichtman is going to get it right, which we don't know yet. TWM03 (talk) 10:45, 2 August 2024 (UTC)

Please stop editing the page to make these into one column. It is clearly not justifiable for Wikipedia to say that the actual winner is likely Kamala Harris (see WP:CRYSTAL), which if you read the column headings is what this is saying. TWM03 (talk) 23:12, 2 August 2024 (UTC)

The unified column say “TBD (Leans Harris)”, which as the state of both the projected and “actual” results as per Lichtman’s views. The columns should be joined whenever there isn’t a textual difference between them. In any case, there’s no need for 2 separate TBD columns. RobertGustafson (talk) 00:16, 3 August 2024 (UTC)

Lichtman has said that the race “leans” toward Harris at the moment, so if we must have 2 columns for 2024, then Projected should be “leans Kamala Harris” and Actual should be TBD. RobertGustafson (talk) 00:19, 3 August 2024 (UTC)

“Likely” was supposed to be “leans”; I have since made the correction. RobertGustafson (talk) 00:23, 3 August 2024 (UTC)

No objection to your 00:19 suggestion. My only objection is to having anything other than TBD in the final column (for which Lichtman's views are irrelevant as it lists the actual winner). Saying the actual winner leans or is likely Harris is a violation of WP:V even if it is Lichtman's opinion. TWM03 (talk) 00:22, 3 August 2024 (UTC)

There is nothing around the table saying that when the columns are joined, it indicates the “actual” winner for a coming election. My point as that we can’t have TBD in both columns as it is textually redundant; by that logic, the column pair for EVERY election should be separate instead of joined. Anyway, the “lean” or “likely” of several keys for 2024 are based solely on Lichtman’s word, and Lichtman’s words regarding the projected winner, while not final, are essentially paraphraseable as “leans Kamala Harris”—as he has said that “a lot would have to go wrong” for her to lose. Yes, this all based on 1 man’s word, but so are all the predictions for the keys for 2024, and it’s his model—making him the preeminent authority. Granted, arguments based on authority have little logical value, but this is Wikipedia, where everything is based on sources, not logic. RobertGustafson (talk) 00:32, 3 August 2024 (UTC)

Quote, "There is nothing around the table saying that when the columns are joined, it indicates the “actual” winner". With respect, if you think this you have not read and understood the table. When the columns are joined it means that the actual and predicted winner are both the named candidate.

Quote, "Yes, this all based on 1 man’s word, but so are all the predictions for the keys for 2024, and it’s his model—making him the preeminent authority." By this argument, the actual winner of the 2000 election would be listed as Al Gore because Lichtman claims that the election was stolen. Have you read WP:V and WP:CRYSTAL? TWM03 (talk) 00:44, 3 August 2024 (UTC)

Since the point of the table is to measure Lichtman’s track record of election prediction, only his opinion of the 2024 lean of keys and outcome matter. Joining the columns acknowledges that the actual winner is undetermined, but parenthetically showing as leaning in Harris’ direction (borne out both by the math of his assessments of the keys and his holistic tentative assessment of the outcome) DOES NOT suggest that Harris is the “actual” winner; no actual election result “leans” one way or the other, and “TBD” makes it clear that the actual winner is undecided. RobertGustafson (talk) 00:44, 3 August 2024 (UTC)

Lichtman’s word is irrelevant when it comes to actual winner, but there is no need for 2 TBD columns. Either Projected should say “leans Kamala Harris” since that’s where he said it essentially LEANS (though he’s yet to make a formal prediction) and Actual should say TBD, or 1 column should say both—combining his tenative prediction with the undetermined outcome. RobertGustafson (talk) 00:49, 3 August 2024 (UTC)

I feel like we’re splitting hairs here! RobertGustafson (talk) 00:50, 3 August 2024 (UTC)

We have one solution that we have said we are both happy with, so that is good. Let's just agree to that. As long as the final column (by which I mean the column whose heading is "actual winner") says TBD and nothing else, I'm happy. TWM03 (talk) 00:54, 3 August 2024 (UTC)

Done. Projected is “leans Kamala Harris” (paraphrased from Lichtman’s remarks on the 2 sources for the column), and Actual is TBD. RobertGustafson (talk) 01:34, 3 August 2024 (UTC)

Lichtman, according to the sources cited in the column, HAS tentatively predicted a Harris win—all the while saying he has yet to make a definitive prediction. Besides, leans or likely presumes an outcome TBD, as those words are never used when the outcome is decided and known. RobertGustafson (talk) 01:37, 3 August 2024 (UTC)

Someone messed up the modern prediction table

Someone moved the entire reception section and bibliography into the 2024 prediction spot. I cant fix it myself right now. Ablevi202 (talk) 22:52, 2 August 2024 (UTC)

There has been a lot of disruptive editing on this page. Should be fixed since I restored a version from earlier today TWM03 (talk) 22:59, 2 August 2024 (UTC)

My bad, as to the refs: It took me a while to figure out the correct source syntax to separate the refs from the table. (BTW, there seem to be limits to how you can edit a table when doing visual editing!) I have tried to keep the Projected and Actual Winner columns for 2024 joined, as they are both essentially “TBD (Leans Kamala Harris)” at present; there’s no point in giving separate apparent winners (projected vs. actual) for 2024 until Allan Lichtnan makes a formal prediction for 2024 or the election selects a 2024 winner. RobertGustafson (talk) 23:51, 2 August 2024 (UTC)

There shouldn’t be 2 TBD columns for the winner

Either the Projected Winner/Actual Winner columns should be “Leans Kamala Harris”/“TBD”, or they should be combined to show “TBD (Leans Kamala Harris)”. A projected winner, as opposed to an actual one, is never TBD!! RobertGustafson (talk) 23:40, 2 August 2024 (UTC)

Of course a predicted winner can be TBD, if the prediction hasn't been made yet. All primary and secondary sources confirm that the prediction hasn't been made yet, so it is TBD. TWM03 (talk) 23:30, 2 August 2024 (UTC)

There still shouldn’t be 2 TBD columns. If the projected winner is TBD, then of course the actual one is TBD too, so there’s no need for a double column, just as there’s no need for 2 columns when the predicted and actual winner match! Aside from the point that a predicted winner can be TBD, I stand by my assertion that there should only be one such column (with the leaning candidate in parentheses). RobertGustafson (talk) 23:40, 2 August 2024 (UTC)

There are two columns in the table, one for projected winner and one for actual winner. I don't care if they are merged and both TBD, or if they are not merged and predicted is TBD (likely Harris), actual is TBD, or if they are not merged and predicted is Likely Harris, actual is TBD. What you cannot have is anything other than TBD in the actual winner column, whether it is part of a merged cell or not. TWM03 (talk) 23:54, 2 August 2024 (UTC)

I have the predicted as LEANS Harris, not Likely anymore. BTW, by your logic that everyone assumes a joined column means both projected and actual are one, everyone understands that a “lean” is a prediction—in which we defer to Lichtman’s word—and therefore, joining the columns does not imply that the actual winner is determined. In any case, joining the columns TECHNICALLY only means that they would say the same thing—which is that the winner is undecided but tentatively (not finally) seen by Lichtman as leaning toward Harris. You are assuming that people who read the article don’t have common sense. RobertGustafson (talk) 01:01, 3 August 2024 (UTC)

I now have Projected as “Leans Kamala Harris” and Actual as TBD. Hopefully this can be the final form until Lichtman makes his prediction. (BTW, note that the use of the word “leans” coupled with the citation makes it clear that the prediction is tenative, not final. Lichtman has made it clear where the race is tentatively leaning, both explicitly [in his interviews and columns] and in terms of the number of false keys.) RobertGustafson (talk) 01:11, 3 August 2024 (UTC)

Short Term and Long Term Economy

Specifically 2016-2020 the long term economy was great Short term there was a small recession through COVID How could they place false on both???

On the other term 2020-2024 we had a major blow out in 2022. Now we are having some work related recession (not stocks recession), people are selling homes and living in mobile houses or renting due to the unaffordability of housing and mortgage

How dares Mr Lichtman claim bad economy 2016-2020 and the reverse 2024? This is not subjective only but biased also. 5.195.85.184 (talk) 10:28, 3 August 2024 (UTC)

The table shows how Lichtman called the keys. Are you suggesting that the article should be changed or just stating your disagreement (because this page is not a forum)? TWM03 (talk) 11:09, 3 August 2024 (UTC)

I never meant that your article is wrong, it is matching Mr. Lichtman's answers.

The article must comply with his predictions, and I think it did.

I think that the keys are subjective here, and Mr. Lichtman has introduced some freedom changing the answer to the key questions.

Recession is defined as 2 quarters back to back falling stock indices, which we never had in 2016-2020, we got a short recession in 2022, which honestly I don't call a real recession.

Recession can be accepted if you, I, he, she and others are unable to find a job, losing their homes.

It is happening now. Half homeless became full homeless specifically in CA south, and I don't remember recruiters calling for job since March 2024. 2001:8F8:1135:6AE6:94DB:D751:372:B899 (talk) 15:59, 3 August 2024 (UTC)

I'm thinking that this section and the below one should be removed as there doesn't appear to be any suggestion for improving the article here. TWM03 (talk) 22:22, 3 August 2024 (UTC)

Probably the only improvement is to give an idea that Wiki can add criticism to the keys modeling saying Mr. Lichtman gave the modeler and the implementer a proven freedom that he can (flip) the Whitehouse pick based on his own subjective argument.

This freedom can change down to Nov while counting the ballots, we can subjectively flip.

Which tells the reader that the model is not accurate at a range of 20-25% of 13 Questions.

I wanted to reach this point interactively with Wiki and this is the end of my arguments. Many thanks for driving the conversation. 2001:8F8:1135:6AE6:C954:AC61:20EF:408A (talk) 14:16, 4 August 2024 (UTC)

I think this is covered in the third paragraph of the introduction -- "Some of the items on the checklist involve qualitative judgment, and therefore the system relies heavily on the knowledge and analytical skill of whoever attempts to apply it."

If you want something to be added to the criticism section, you would need to find a reliable source stating that the keys have been criticised on this basis. TWM03 (talk) 08:33, 5 August 2024 (UTC)

Incumbency Factor

Eventually incumbency is very beneficial. But looking at the incumbent President 2024 he had no chance defeating the challenger. The party made a last minute change and forced the incumbent to step down.

Now as per the model, the replacement is having the privilege according to the keys.

That makes the incumbency looks like -1 or a negative key. Either the modeling is wrong or the subjectivity of the model was applied wrong. 5.195.85.184 (talk) 10:54, 3 August 2024 (UTC)

Economy Factor

Someone changed the Short Term Economy to "Toss Up" based on the current stock market issues and potential recession. However Lichtman himself has not said anything on the matter and the article is meant to list the keys per Lichtman's predictions not how we view certain factors. Shouldn't the key remain true until Lichtman comments on it? — Preceding unsigned comment added by 92.15.6.130 (talk) 17:39, 5 August 2024 (UTC)

I agree. I honestly think the page should be locked until the final prediction is made. 155.190.22.1 (talk) 18:31, 5 August 2024 (UTC)

People talking about the Sahm rule don't know what they're talking about. Prof. Sahm has already come out and said recent numbers don't indicate a recession because A) They are very slight and B) there are unique circumstances at play.

Also, the stock market is not crashing or anything so dramatic. Overvalued stocks suffered a market correction which is within normal trading margins. In order to be a "crash" the market would need to dip -5% over 1 to 2 days, or have a full week of substantive negative slide. Keep in mind the dow is still +2-3% over the past 6 months. 107.220.118.81 (talk) 19:40, 5 August 2024 (UTC)

Incorrect Allan Lichtman prediction.

This article states that Lichtman incorrectly predicted the 2016 election when in fact he received massive media coverage for being one of the people to have correctly called Donald Trump to win. What I assume the writer is instead referring to is the 2000 election where he incorrectly predicted a victory for Al Gore and later defended the 13 Keys model by claiming he was only predicting the popular vote (which Al Gore won) as a proxy for the electoral college. He claims to have since altered his approach and correctly called the winner of the 2016 election. He also claimed in a 2024 livestream (Will the Prisoner Exchange Help Dems in November??? | Lichtman Live #63) that his prediction was right and the 2000 election was stolen from Al Gore and misattributed to Bush. 2A00:23C5:11E:F901:945F:1702:FD34:FB31 (talk) 16:08, 8 August 2024 (UTC)

(same user) To be honest, Lichtman looks like a bit of a hack, his model is only supposed to predict the popular vote as a proxy and he was incorrect in that regard by calling for Donald Trump in 2016. I haven't seen any evidence that he's factored the electoral college into his approach since his 2000 incorrect prediction although I haven't read his publications. Joko246 (talk) 16:19, 8 August 2024 (UTC)

This was addressed a little more thoroughly in the "2016 Election Confusion" section above on this talk page. Some recent reporting and the location of the October 2016 paper which contradicts his post-2016 narrative make the current language justifiable and, if anything, pretty light on his record. Caraturane (talk) 01:10, 9 August 2024 (UTC)

Thanks, the page has changed since I wrote that maybe an edit was reversed. What's up there now is good. 2A00:23C5:11E:F901:945F:1702:FD34:FB31 (talk) 09:42, 9 August 2024 (UTC)

His model was always supposed to predict the popular vote. Even in the retrospective applications of his keys, it predicts popular vote, not electoral college outcome. Trump lost the popular vote in 2016 despite Lichtman giving him enough keys to win. He won because of the electoral college. 216.174.98.252 (talk) 19:39, 10 August 2024 (UTC)

Lichtman is being dishonest by changing what his model predicts without actually changing the model. Also, he said his model predicted the popular vote as late as September 2016 and it wasn’t “corrected” as predicting until after the 2016 election. 216.174.98.252 (talk) 19:41, 10 August 2024 (UTC)

Third Party Key in 2016

Allan Lichtman's book Predicting the Next President: Keys to the White House 2020 Edition, states "In an interview with the Washington Post on September 23, 2016, I predicted that Donald Trump would win the presidency in November, after a sixth key turned against the incumbent Democrats, revising my earlier indication of a likely Democratic win. The sixth key turned because Libertarian candidate Gary Johnson exceeded 10 percent in more than one poll, which, according to the decision rule, turns this key in advance of the election.". So why is the third party key for the 2016 election indicated as true? The official Keys prediction for 2016 should not be published according to how Wikipedia editors interpret the keys, but according to how Lichtman himself turned the keys in 2016. It doesn't matter that Gary Johnson didn't win 5% of the vote in 2016, what matters is that Lichtman himself turned the key - that is what should be in the official predictions table, not the opinions of anyone other than Lichtman. EdDwardian (talk) 10:31, 10 August 2024 (UTC)

If I recall correctly, Lichtman changed his mind on the third party key after Gary Johnson got less then 5% of the vote. He left the Primary Contest key undetermined, which he then claimed was determined after the election. However, his reasoning for leaving it undetermined was dubious. Here it is right here:

A)Unlike in past party nomination struggles, he did not take his candidacy to the convention

B) “The spectre of a Trump victory has United Democrats regardless of the Sanders/Clinton contest 216.174.98.252 (talk) 19:38, 10 August 2024 (UTC)

Could I have the source? It is definitely possible that he changed his mind on the third party key after the election but I found this highly unlikely given that the 2020 edition of his book states that the key was false in 2016. EdDwardian (talk) 01:27, 11 August 2024 (UTC)

Need more information on his retrospective application of the keys

Someone should add information about how Allan Lichtman retrospectively applied the keys. Did he develop the keys first, then retrospectively apply them, and end up with a perfect popular vote predicting system? Or did he try many different combinations in order to find the right combination that worked? Did he do his retrospective applications before he decided to use his keys? If so, then there’s a real possibility that this is data mining. More information on this would be appreciated. 216.174.98.252 (talk) 19:33, 10 August 2024 (UTC)

Lichtman called for page to be edit

As I noted on the article about Lichtman, he called for this page to be edited on his Youtube livestream tonight, and the most recent edits to this page roughly correspond with the stream. Not sure how this is usually handled but seems like an argument for those edits to be reverted.— Preceding unsigned comment added by AmeriMike (talk • contribs) 02:31, 16 August 2024 (UTC)

This is a conflict-of-interest edit and raises other policy violations. I have reverted these edits including those on the Allan Lichtman page and informed Wiki admins to see about getting more protections on these pages. Caraturane (talk) 03:21, 16 August 2024 (UTC)

DNC Protests

Even a likely repeat to the 1968 Democratic National Convention riot activity could only be too localized to turn the key false since national conventions focus on one city, which is unlikely to flip the social unrest key. Is that correct? 2604:3D09:6578:4800:20CB:2FE4:261E:4BB5 (talk) 21:53, 20 August 2024 (UTC)

No, to add this you would need at least the following: (a) a reliable, secondary source stating that there will likely be a repeat of the 1968 riot activity at the DNC, and (b) a source where Lichtman states that even if this happened it would not flip the social unrest key. I believe that neither of these conditions has been met.

Moreover, there has been a lot of non-encyclopedic speculation in the table including details that are just not necessary for a Wikipedia article. The final prediction is getting made in 2-3 weeks anyway; I see no problem with waiting until then (or at least, until after the DNC protests happen or don't happen) to decide whether to add details like this. TWM03 (talk) 23:33, 20 August 2024 (UTC)

I was just worried. 2604:3D09:6578:4800:20CB:2FE4:261E:4BB5 (talk) 00:02, 21 August 2024 (UTC)

Worried about the stability of America herself. 2604:3D09:6578:4800:20CB:2FE4:261E:4BB5 (talk) 00:24, 21 August 2024 (UTC)

No third party is fully true now

I just wanted to give you all an update that RFK Jr. plans on dropping out and endorsing Trump. NesserWiki (talk) 22:37, 21 August 2024 (UTC)

We'll wait for Lichtman to update where the Keys are leaning, though. NesserWiki (talk) 22:39, 21 August 2024 (UTC)

Should the foreign policy success key be undecided instead of leans true?

I mean, Lichtman did not explicitly say that this key leans true. 100.16.156.64 (talk) 11:53, 22 August 2024 (UTC)

Changed. TWM03 (talk) 20:27, 22 August 2024 (UTC)