I first observed an arithmetic Republican lean in Emerson College (“Emerson”) polling in November 2019. Specifically, Emerson College polling of hypothetical 2020 matchups between leading Democrats (Joe Biden, Bernie Sanders, Elizabeth Warren) and President Donald J. Trump had shifted 4.7 percentage points (“points”) Republican since September 1, using the Democratic percentage minus the Republican percentage (“DR margin”). New polling allowed me to update this analysis five days later.
I again updated these analyses in August and October 2020; the latter compared Emerson’s Republican lean to the Democratic lean of Quinnipiac University’s (“Quinnipiac”) polling, arguing you could determine the “true” final DR margin in an election by simply averaging Emerson and Quinnipiac polls.
Except, that was not true. While Quinnipiac overestimated the final national Biden-Trump DR margin (4.5 points) by 5.9 points, Emerson underestimated that margin by just 0.7 points; DR margins were calculated using my weighted-adjusted polling average (WAPA) methodology. In fact, Emerson not only came the closest of all pollsters with at least a B+ rating from FiveThirtyEight.com, they were the only one not to overestimate Biden’s final nation margin over Trump. I concluded every other pollster had once again underestimated Republican turnout with Trump atop the ballot.
After the 2020 elections, FiveThirtyEight.com recalculated their pollster ratings. Emerson remained at A-, along with a historic Democratic (!) bias of 0.8 points based upon 201 state-level polls. Quinnipiac was upgraded to A-, with a Democratic bias of 0.5 points.
Meanwhile, in the summer of 2021, I began to track “generic ballot” polls. These ask some variation of “If the election in your district were held today, would you vote for the Democratic candidate, the Republican candidate, some other candidate or are you not sure?” Some versions query which party you would prefer to see have the majority in the United States House of Representatives (“House”), but the idea is the same: which party is leading nationally, irrespective of candidate specifics.
As of August 30, 2022, I have compiled 384 such polls, dating from February 2021. On average, Republicans lead Democrats by 0.6 points (43.3% to 44.0%, with rounding), though if you examine only the 10 most recent polls, Democrats lead by 1.3 points (45.1% to 43.7%, with rounding). In fact, Democrats have mostly led since the last week of July, based upon rolling 10-poll averages.
Emerson has released seven generic ballot polls thus far in 2022, with the most recent conducted August 23-24, giving Republicans a one-point lead, 45% to 44%; adjusting for historic lean, this equates to a 1.8-point lead – 1.2 points more Republican than the overall average, and 3.4 points more Republican than the 10 most recent polls not conducted by Emerson. Overall, Emerson has estimated Democrats trail Republicans 42.2% to 46.8%.
Table 1: Generic Ballot Polls Since February 2022 – Emerson College vs. All Others
|Poll dates||Adjusted Emerson DR margin||All Others DR margin*||Difference|
*Using WAPA calculated from the five polls conducted before and the five polls
conducted after, based upon field date midpoint
As Table 1 shows, Emerson polls suggest a political environment 3.5 points more Republican than do all other pollsters over a similar time frame.
Does this same Republican lean exist for other 2022 elections? Since November 2021, Emerson has conducted 19 2022 United States Senate (“Senate”) and gubernatorial race polls (Table 2):
Table 2: 2022 Senate and gubernatorial polls – Emerson College vs. All Others
|Election||Poll month (2022)||Adjusted Emerson DR margin||All Others DR margin||Difference|
|GA Senate (2)||April, August||-3.3||1.3||-4.6|
|NC Senate (2)||April, May||-7.8||0.4||-8.2|
|Total Senate (9)||May-August||4.0||7.2||-3.2|
|GA Governor (2)||April, August||-6.6||-5.6||-1.0|
|Total Governor (10)||Feb-August||3.4||5.6||-2.2|
*Using average of polls assessing Democrat Maura Healey against Republicans
Geoff Diehl and Chris Doughty
This VERY back-of-the-envelope math suggests Emerson has an average Republican lean of 3.2 points in the seven Senate races it has assessed, two of them twice (Georgia, North Carolina), compared to an average Republican lean of 2.2 points in the eight governor’s races it has assessed, one twice (Georgia). Looking only at the three races Emerson has assessed twice, the Republican lean jumps to 4.6 points – but with the strong caveat Republican Ted Budd was performing much better against Democrat Cheri Beasley (R+3.6 excluding Emerson) earlier in the year than he has in recent months (D+1.3).
Still, these Senate and governor’s race lean estimates are not substantively different from those for the generic ballot – converging on a Republican lean of roughly 3 points. Curiously, this is about what I would expect if the true undecideds broke about 2-1 Republican in all of these races.
There is one final way to assess Emerson’s possible Republican lean in 2022 – examining hypothetical matchups between Biden and Trump in 2024, both nationally and by state.
Table 3: Hypothetical 2024 Biden vs. Trump polls – Emerson College vs. All Others
|Election||Poll month (2022)||Adjusted Emerson DR margin||Actual 2020 DR margin||Difference|
|Total National (7)||Feb-August||-2.8*||4.5||-7.9|
|Total State (5)||May-August||-3.8||3.2||-7.0|
*Weighted by month, with August=64, July=49, June=36, etc.
The seven hypothetical national 2024 Biden-Trump matchups assessed by Emerson in 2022 imply an average Republican shift of nearly eight points from 2020. The sparse state-level data, meanwhile, suggest a Republican shift of 7.0 points. Overall, these 12 polls point to a 7.5-point shift toward Republicans – consistent with Democrats losing the generic ballot by 4.4 points, almost exactly what Emerson polls suggest.
Sitting at the other end of the “outlier” spectrum, meanwhile, is YouGov. This B+ pollster (lean D+0.7) has conducted 87 generic polls since April 2021, 69 of which are weekly tracking polls for The Economist. Across all of these polls, YouGov gives Democrats a 3.2-point lead (43.2% to 40.0%), albeit with a higher percentage choosing neither Democrats nor Republicans – 16.8% vs. 12.1% for all other polls. As with Quinnipiac, the “truth” appears to lie between You Gov and Emerson: R+0.7.
I draw no conclusions as to which pollster – if either – is “correct.” All I will say is that there is a reason I assess all elections three different ways: assume polls are unbiased, on average (R+0.6, using the generic ballot); assume polls consistently overestimate Democratic strength by three points (R+3.6, closer to Emerson); and assume polls consistently underestimate Democratic strength by three points (D+2.4, closer to YouGov).
We shall learn which method was the most accurate in about 10 weeks.
Until next time, please wear a mask as necessary to protect yourself and others – and if you have not already done so, get vaccinated against COVID-19! Also, if you are not already registered to vote, please do so immediately. And if you like what you read on this website, please consider making a donation. Thank you.