Talk:List of election bellwether counties in the United States

Semi-protected edit request on 14 November 2020
Otsego County, NY has correctly predicated the Presidential Election every year since 1976, minus 2020.

Cortland County, NY has correctly predicated the Presidential Election every year since 1976, minus 2020, losing by a margin of 421 votes in favor of Former President Donald Trump. 2604:6000:1412:4222:3C37:19A2:4343:99E (talk) 18:52, 14 November 2020 (UTC)
 * Red information icon with gradient background.svg Not done: please provide reliable sources that support the change you want to be made. &#8209;&#8209; El Hef  ( Meep? ) 22:37, 14 November 2020 (UTC)

2020 election exceptional?
This article makes a claim in the lead which is proven wrong by the article itself. It currently reads "in the 2020 presidential election, almost all counties previously considered reliable indicators of eventual success in the election did not continue their streaks as bellwethers, voting for Donald Trump as opposed to the projected winner of the election, Joe Biden." The claim is supported by a wall street journal article which says the same thing. However, our article goes on to list 43 counties which deviate in 1 election, 18 of which voted for Trump and 25 of which voted for Biden. So we have a reliable source claiming one thing, but our very article demonstrates something else. Can we remove this under WP:BLUESKY? Awoma (talk) 10:47, 19 December 2020 (UTC)
 * The key word is "previously". The truth or falsity of the statement you removed from the article is determined by counting using the lists of counties with 0, 1 or 2 recent deviations prior to the 2020 election.  You appear to be opposing the statement by counting the revised lists after the 2020 election, but those are not the relevant lists to test the claim and it is circular to use them to disprove a change (since that change has been used to adjust the data in the test) . 73.89.25.252 (talk) 01:36, 22 December 2020 (UTC)
 * Using the lists in the article, there are 106 counties which (since 1980) had predicted at most one election wrong before 2020, and 81 of 106 (about 80 percent) voted for Trump in 2020. And of the 19 "bellwether" counties that had predicted all elections correctly since 1980, 18 of 19 voted for Trump in 2020; this confirms the claim in the Wall Street Journal article that was cited as the source for the material removed by .  WSJ wrote:


 * From 1980 through 2016, 19 of the nation’s more than 3,000 counties voted for the eventual president in every election. Only one of them, Washington state’s Clallam County, backed President-elect Joe Biden last week.


 * Other counties that had been bellwethers all the way back to the 1950s ended their runs by backing President Trump instead of the Democrat.


 * Hence, whether one uses only the bellwethers with a perfect record through 2016 or the broader list of almost-perfect bellwethers with at most one error, the claim in Wall St Journal that the bellwethers' track records were decimated is supported by the WP:OR lists so heroically compiled into this article.  How unusual this decimation is compared to what happens in a typical election, is a question for further analysis, but as a factual assertion the material from WSJ is correct.  73.89.25.252 (talk) 07:07, 22 December 2020 (UTC)
 * Yes. If we consider those counties which had at most one deviation between 1980 and 2016, the majority voted for Trump. As you say, how unusual (and therefore notable) this is is unclear. Awoma (talk) 08:42, 22 December 2020 (UTC)
 * The lists are OR, and may have been compiled for political purposes (such as diluting the Trumpist claims by ostensibly "updating" things through 2020, thus obscuring the decimation pattern; certainly there was a flurry of edits once the business about bellwethers began appearing online after the election). The reduced data set with 19 counties did get some notice in national media, as in the WSJ and some television coverage, so that's the apparently notable part for 2020.  From the data in the article, it looks like in 2016 Trump won a slight majority of the bellwethers with perfect records prior to that year, which seems like what one would expect in a model where these middle of the road counties are somewhat biased toward the national trend and the rest is luck (for the perfect predictors list; the broader list including the almost-perfects may be more representative).  But under that type of model one would not expect 18 of 19 and 80% of the almost-perfects to go to the loser in 2020. 73.89.25.252 (talk) 09:45, 22 December 2020 (UTC)
 * Indeed. That model doesn't work, as explained in the WSJ article. Counties shifting from Trump in 2016 to Biden in 2020 are few and far between. Awoma (talk) 10:14, 22 December 2020 (UTC)

Previous version of this article
This article previously listed counties that had gone with the POPULAR VOTE WINNER in every election since 1980. This was useful information and should be re-added.Amyzex (talk) 15:00, 25 June 2021 (UTC)

Unclear article content
Someone needs to add an explanation of what the bold font means in the "Deviations in two elections" section, or otherwise remove that formatting. (u t c [ m] [ l] ) 🔒 ALL IN 🧿 21:46, 15 September 2023 (UTC)


 * I agree. Eyeballing it, it looks like years Republicans won are in bold.
 * It has the feel of a data collection project for partisan purposes. I'm not sure why so much of the article is taken up with a list of "deviations in two elections" since 1980. Lana Del Taco (talk) 16:59, 13 March 2024 (UTC)

What is the purpose of this article?
This article is little more than a collection of data related to all-or-nothing bellwethers using 1980 as a base year. Is there any good reason for focusing solely on how counties have voted since 1980 or to have a huge list of counties that have deviated from the winner 1 or 2 times since 1980? It feels like this article only exists in its current form because of the claim of Trump supporters that Biden winning only 1 of 19 counties that voted for every winning presidential candidate from 1980 to 2016 is indicative of fraud.

Additionally, this article exclusively covers all-or-nothing bellwethers, stating the “strongest bellwether counties are those that back the winning candidate in all elections,” but research has found such bellwethers are not good predictors of future results. The article doesn’t also discuss any research on bellwether districts and doesn’t discuss barometric or swingometric bellwethers at all.

I'm not sure this page should even exist when the politics section of Bellwether has more useful information on the topic. Lana Del Taco (talk) 15:58, 13 March 2024 (UTC)