Talk:The Keys to the White House/Archive 4

Archive 1

Archive 2

Archive 3

Archive 4

Lichtmann's rebuttal video and the popular vote statement

In Lichtmann's rebuttal video dated 17/10/24, he argues that a statement in his 2016 paper has been taken out of context and misunderstood as him claiming that the model predicts the popular vote rather than the electoral college tally. He claims that media articles have promoted this misunderstanding.

That statement from 2016 reads as follows: "As a national system, the Keys predict the popular vote, not the state-by-state tally of Electoral College votes. However, only once in the last 125 years has the electoral college vote diverged from the popular vote."

He claims that he was simply making the point that the model in no way looks at individual states. To be fair to him, this is a distinction between the 13 Keys model and polling that he seems to be in the habit of making in media interviews, often said in a similar fashion but with more clarity. Can we once and for all discuss whether this statement clearly merits him stating that his model only predicts the popular vote and put this to bed.

For my two cents it seems he should have no idea whether the model predicts the popular vote or electoral college since the split never occurred in his dataset-- the only thing he can logically derive from his dataset is that his model should predict election winners. This is a comment speculating on his thought process and not my own judgement.2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 01:08, 20 October 2024 (UTC)

Having re-read the statement, I'm conflicted due to the fact he starts the sentence with "As a national system" as well as using the otherwise defunct adjective "state-by-state". He could be saying: "The model predicts the popular vote (not exclusively) and does not conduct a state-by-state tally of the electoral college"

It is then confusing as to why he feels the need to say the electoral college rarely diverges from the popular vote as if he's arguing in favour of the value of his model in the face of a potential liability.

Perhaps I'm tying myself in knots, I'll let people tell me but a big part of why I'm raising this is his media interviews. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 01:23, 20 October 2024 (UTC)

I believe you are referring to this video clip:

https://www.youtube.com/watch?v=_Af3hKnrexs.

The job of the Wikipedia is to summarize key points, not to take sides re: WP:RS. If there is a point of controversy, then an article lists the various arguments in a controversy section. This video raises concerns about two areas: WP:UNDUE or undue weight and Wikipedia:Cherrypicking - both areas that should be avoided in WP:BLP. The subject article is not my area of expertise, I work on various WP:BLP (biography of living persons articles), so perhaps other editors who are familiar with the subject can weigh in.-Classicfilms (talk) 01:45, 20 October 2024 (UTC)

Great thank you but surely Wikipedia must consider some things objective fact? I only say this because it is Apprentice57's argument in the RfC. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 01:51, 20 October 2024 (UTC)

Apologies, I'm not sure I understand your question entirely, feel free to rephrase it if you like- The Wikipedia is generally guided by WP:NPOV, which means looking at the wide scope of a topic- that is why I raised WP:UNDUE and Wikipedia:Cherrypicking. The goal is to avoid WP:OR, or original research, which can happen if a series of facts are strung together. I'm not saying that is happening here, I'm trying to respond to your overall questions. -Classicfilms (talk) 02:05, 20 October 2024 (UTC)

You ask a good question, and over the years Wikipedia has developed its answer. As Classicfilms said, the policy is set out in Wikipedia:Neutral point of view. Wikipedia generally does not take sides, but fringe viewpoints need not be given equal attention. One example given is that "the article on the Earth does not directly mention modern support for the flat Earth concept, the view of a distinct (and minuscule) minority; to do so would give undue weight to it." In that sense, the spherical Earth is treated as an objective fact.

That doesn't mean, however, that, for every dispute, Wikipedia editors consider the evidence, decide which opinion is correct, and present it as fact. Actually, the flat-Earth case is the exception. The general rule from WP:NPOV is:

Avoid stating seriously contested assertions as facts. If different reliable sources make conflicting assertions about a matter, treat these assertions as opinions rather than facts, and do not present them as direct statements.

Here, there are seriously contested assertions about the 2016 prediction. Apprentice57's argument in the RfC is that one side is correct and should therefore be presented as a direct statement. (Version 1 flatly asserts: "Lichtman's model incorrectly predicted the popular vote outcome.") That would violate the NPOV policy. JamesMLane t c 02:19, 20 October 2024 (UTC)

Okay thank you both for clarifying and doing so accessibly. Should this specific controversy on the language and intended meaning of the statement be included in the article? It appears to be the main source of this dispute and the central premise behind several cited articles that have fuelled it. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 02:51, 20 October 2024 (UTC)

In my opinion, yes, this controversy is notable and should be included in the article. Version 2 presents both sides without adopting either. JamesMLane t c 03:05, 20 October 2024 (UTC)

Yes, agree with JamesMLane on all points above. You might want to also look at this article, WP:NOT or "What the Wikipedia is Not," which may further clarify your questions. -Classicfilms (talk) 03:20, 20 October 2024 (UTC)

Remember, the reason why the October 2016 paper is so crucial, and a smoking gun as to the fact it's a popular vote model here: there was no dispute! Lichtman, and his critics (like Nate Silver) all agreed it was a popular vote mode. That is why the dispute is a based on lichtman taking a fringe position.

That it has come under dispute post facto is irrelevant to a prediction. Predictions are all about what is said and claimed beforehand, not after. And Lichtman would have a motive in being dishonest here, as it buoys his reputation to be correct. Apprentice57 (talk) 12:48, 20 October 2024 (UTC)

The only neutral point of view is what Lichtman said about his predictions before November 2016. Those statements where he said what his system predicted and what he predicted are very clear and they directly note that he was only predicting the popular vote and say it was not predicting the Electoral College. This is a neutral POV. Caraturane (talk) 17:03, 20 October 2024 (UTC)

It is an objective fact that Lichtman said his system only predicts the popular vote. There are years and years of him saying so. Tomcleontis (talk) 16:07, 21 October 2024 (UTC)

This isn't necessarily of importance to the RfC/his article, but I think reading one of his older intro chapters in his book (which he republishes every 4 years roughly) is helpful. I did so with his 1990 book on the internet archive (importantly, old enough to be before 2000), my conclusion is that he recognizes the keys are a popular vote model (it states so explicitly more than once), but that was seen as sufficient to also predict the winner: https://www.reddit.com/r/fivethirtyeight/comments/1dc7wiy/allan_lichtmans_the_thirteen_keys_were_incorrect/ Despite the age of the 1990 book, the text of the keys has not changed.

His October 2016 paper is of importance because it establishes that there was no dispute to the keys predicting the popular vote on election's eve, it repeats most of the language from his books but shorter owing to the format. Apprentice57 (talk) 12:54, 20 October 2024 (UTC)

Thanks for this but I must say that these statements are never exclusive and it does seem possible he's making the point that the keys are a national system rather than working state-by-state. I certainly held your view once and I get it but this could a case of miscommunication and clumsy language. I think this is supported by it being simply not rational for him to make this distinction, there's no basis for it in his dataset and these statements as you and almost everyone have read them don't make sense. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 14:55, 20 October 2024 (UTC)

Strong disagree! But I appreciate the polite pushback. Apprentice57 (talk) 00:05, 21 October 2024 (UTC)

The split occurred in his dataset in 2000. He took credit for 2000 (like in this New York Times video) by saying he only predicted the popular vote. Caraturane (talk) 16:28, 20 October 2024 (UTC)

He was of the opinion that Al Gore won in 2000, you can see his submission to the federal commission cited in this article, it's also something he has restated a multitude of times in his YT videos and livestreams. It was the NYT narrator that told us about the 'split' not Lichtmann. All Lichtmann said was that he correctly predicted the popular vote which he did. Up until 2000, his model had correctly predicted the popular vote and electoral college. Because of his opinion on the 2000 election, in his view, he had also correctly predicted the electoral college.

We have to remain mindful of the context in which he made these statements-- to his beliefs, the popular vote had still not diverged from the electoral college until 2016. Given that, I don't think it's right to represent him as a liar because of this. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 16:41, 20 October 2024 (UTC)

No one is representing him as a liar (though it's not for nothing he has been caught being inconsistent and not telling the truth about some things like saying he switched to only predicting the winner after 2000, this is provably not true). It's notable that even in 2012 he used the same "the Keys predict the popular vote, not the state-by-state tally of Electoral College votes" line and even put in plain English in 2008 that a Keys “Win indicates the popular vote outcome for the party in power.”

This article is helpful to provide sources, I have been using it as a reference to find links that explain.

Caraturane (talk) 17:01, 20 October 2024 (UTC)

Thank you, the 2008 paper is far clearer although his subsequent assertion that the point he makes in that statement is to distinguish between his model and polling calls into question if not this interpretation then the way we treat the statements in my view. It still raises the confounding question of why he would decide that his model only predicts the popular vote, this categorically cannot be derived from his dataset of elections from 1860-1980 that formulated the model (later elections constitute an application of the model). Again his use of the adjective "state-by-state" in most of these statements potentially supports this.

I think to take the position we're currently taking in light of WP:NPOV, we need undeniable proof that his system does not predict the electoral college. The fact that he often links his model to the popular vote may not be exclusive. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 17:21, 20 October 2024 (UTC)

The statement "the Keys predict the popular vote, not the state-by-state tally of Electoral College votes" is a direct assertion it does not predict the electoral college. He also wrote many times after 2016 that he made a change to his system to make it now just predict the Electoral college (though there is not documentation of this change before November 2016, only after the call was made "wrongly" in 2016, for example see this NPR article ). Caraturane (talk) 17:25, 20 October 2024 (UTC)

With all due respect, this NPR article does not evidence a 'revision'. You're of the view that the statement I highlighted in the original post is incontrovertible evidence of Lichtmann saying that the model does not predict the electoral college, I'm arguing that we should be less concrete in our interpretation. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 17:33, 20 October 2024 (UTC)

He uses the word "not," I don't know how much clearer that could be.

And sorry, I should have been more clear: I was using the NPR article as an example of him not mentioning he changed the system. He has written he has many times after the 2016 election, for example: after saying in 2016 multiple times he only predicted the popular vote, Lichtman wrote in a post for the Harvard Data Science Review in 2020 which said:

"In 2016, I made the first modification of the Keys system since its inception in 1981. I did not change any of the keys themselves or the decision-rule that any six or more negative keys predicts the defeat of the party holding the White House. However, I have switched from predicting the popular vote winner to the Electoral College winner because of a major divergence in recent years between the two vote tallies."

Caraturane (talk) 17:36, 20 October 2024 (UTC)

I understand but you must see my point in the original post and first comment-- I'm not arguing that this is a reasonable interpretation for most readers, only that there's a good chance that it reflects reality.

The language he uses in his prediction proclaimations is part of his Keys system. He may have been under the belief prior to 2016 that the popular vote and electoral college was extremely unlikely to split due to the reasoning I set out above and subsequently fixed his language to be more precise. He explicitly says that he had not revised the keys themselves. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 17:46, 20 October 2024 (UTC)

At least a good enough chance to invoke WP:NPOV. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 17:47, 20 October 2024 (UTC)

I strongly disagree. This would take an extraordinary leap in usage of language. He said over, and over, and over, that he did not predict the Electoral College, that he only predicted the popular vote, and even referenced past elections prior to his life to show the keys only predict the popular vote (he wrote in 2004, for example: "As a nationally-based system the Keys cannot diagnose the results in individual states and thus are more attuned to the popular vote than the Electoral College results. The 2000 election, however, was the first time since 1888 that the popular vote verdict diverged from the Electoral College results. And the Keys still got the popular vote right in 2000, just as they did in 1888 when Democrat Grover Cleveland won the national tally but lost in the Electoral College to Republican Benjamin Harrison and in 1876 when Democrat Samuel Tilden won the popular vote but lost the Electoral College vote to Republican Rutherford B. Hayes.").

The word "not" means they do not. If this is really what the matter of dispute comes down to then I think we have a very clear cut and dried result. Caraturane (talk) 17:49, 20 October 2024 (UTC)

But he's still making that statement potentially with the subtext that his model is a 'national system' rather than 'state-by-state'-- isn't it strange that in every statement we're considering, he uses this same language that would otherwise be irrelevant to the point you (and to be fair the vast majority of people) interpret him as making?

I'm not contesting the meaning of the word 'not' but the two clauses that come before it and the subtext of the statements. I sympathise with your viewpoint but I think it's best for other editors to weigh in, it seems we've made our arguments as best we can. 2A00:23C5:11E:F901:25AE:ADFB:9C0C:A589 (talk) 18:04, 20 October 2024 (UTC)

Is this really how far we've come? "Not" means not. This cannot seriously be the argument, right? Tomcleontis (talk) 16:06, 21 October 2024 (UTC)

No it's absolutely not the argument I'm making and I've just explicitly stated that in the comment you're replying to. 2A00:23C5:11E:F901:C0EE:D9DA:EB5C:31AC (talk) 16:12, 21 October 2024 (UTC)

Until someone can identify any area where Lichtman said he was changing what the keys predicted before the 2016 election, I believe this is a circular argument. There is only evidence that he switched to predicting the Electoral College for 2020 and now 2024, everything before November 8, 2016 cleanly puts that the keys predict the popular vote. Caraturane (talk) 17:17, 21 October 2024 (UTC)

I don't disagree that it is a valid viewpoint and the one that's most reasonable but it is an interpretation of the facts. I've suggested that we present the core facts themselves.

I will actually step back from this page now, I recognise this is frustrating for longtime editors. 2A00:23C5:11E:F901:C0EE:D9DA:EB5C:31AC (talk) 17:56, 21 October 2024 (UTC)

I concur that Lichtman's words through 2016 unambiguously point to him predicting only the popular vote.

For example, see this similar section from his book "Predicting the Next President: The Keys to the White House 2016":

""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.

In three elections since 1860, where the popular vote diverged from the electoral college tally—1876 (when Democrat Samuel J. Tilden won the popular vote, but lost in the electoral college to Republican Rutherford B. Hayes), 1888, and 2000—the keys accurately predicted the popular vote winner. 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."

I actually don't know what more you could possibly add to make this clearer. Lichtman (1) states that the Keys ONLY predict the national popular vote, (2) indicates that he's fully aware that the EC & popular vote usually but don't always correlate, and (3) specifies that when the EC & popular vote are split, the Keys predict the popular vote winner.

This isn't remotely out of context, either - the very next paragraph reaffirms that the keys predict the popular vote results (no mention of EC), and indeed the entire book's methodology uses popular vote tables and reasons from them. Hangways1 (talk) 09:36, 1 November 2024 (UTC)

This is a fair conclusion, I thought it was important to make the case for Lichtman's argument in an effort to demonstrate that his record was somewhat disputed.

It's tricky to cite Lichtman's view especially when he reliably behaves in bad faith (I've read his attempted rewrite) but I do think it's worth providing in order to maintain that neutral balance, just sticking to a presentation of the most basic facts.

I'm very satisfied with what's on the page now, especially in that it explicitly refers to the referenced sources rather than adopting their argument while leaving it up to the reader to weigh the merits of Lichtman's testimony. 2A00:23C5:11E:F901:E91A:AC65:47F4:CE9F (talk) 01:57, 17 November 2024 (UTC)