POLL RESULT SUMMARY (SEE FULL RESULTS & EXPLANATIONS BELOW)
Poll
Date
Sample + Party Distribution
Biden
Trump
Pollster Avg
Ref's Avg
*Polls that do not include internals are excluded (Survey USA, YouGov LV model). Quinnipiac is excluded due to internals that bare no relation to reality.
D+5 is too high. Assuming 34D/30I/33R. Data: Biden 83-13D, 40-36I, Trump 83-11R. Adjusted 43.9 Biden to 42.6 Trump. Allocated to 95% of vote, Biden 48.2 to Trump 46.8.
The D+3 sample oversamples Democrats. Adjusting to 34D/30I/33R. According to the data, Biden won D's 93-5, I's go Trump 46-44 and Trump won R's 90-5. Applying the ratios to the more likely turnout, Biden receives 47.8 and Trump 43.8. Biden receives 46.5 and Trump 45.5.
I looked at all the internals provided behind the pay wall. I can't reveal them, but the numbers look very solid. Rasmussen seems to be catching a trend that seems visible in daily life but doesn't show up much in poll, a large group of Dems tired of the wokeness of the new Dem party.
1499RV, For RV41D/26I/28R (Internals unavailable for LV model, excluded)
49
41
x
Political Ref's Take on the above poll
The D+13 sample is unprecedented in the modern era, indefensible. The LV model does not offer internals. Using RV model. Assuming 34R/30I/33D. Data says Biden 90-4 D's, 44-41 I's, 87-4 R's. An RV filter will apply to I's. Adjusted, Biden 45.1 to Trump 42.4.
Adjusting to 34D/30I/33R. Biden did 2 points better across party. Ratios inferred, Biden won D's 92-4, I's 48-45, Trump R's 91-5. Applying ratios to more likely turnout, Biden receives 47.3 to Trump 44.9.
Adjusting to 44D/11I/43D and applying RV filters to D and I columns. Ratios, 97-2 (95-4), 50-40 (46-40), 93-5. Applying ratios to more likely turnout, Biden receives 49.0 to Trump 46.2.
The D+4 is too much. Adjusting to 45D/10I/44R. Voter ratios according to data, Biden 94-4 D, 54-32I (47-39), Trump 92-5R. Assuming a 95% turnout among RV's, which is not going to happen. Applying an RV filter to I's. After adjusting, Biden receives 49.2 to Trump 46.2.
Data says Biden 92-4 D's, Biden 48-21 I's, Trump 88-6 R's. This is a leaners turnout, 43D/13I/36R. Adjusting to 43D/13I/42R. Applying ratios, 48.3 Biden and 41.4 Trump.
The D+8 sample oversamples Dems. Adjusting to 35D/26I/34R. Biden leads by 8 with an +8 Dem turnout, so they must have done evenly across party. Assuming Biden won D's 91-5, I's split 45-45 and Trump won R's 91-5. Applying ratios, Biden receives 45.3 to Trump 44.7.
Will adjust to 40D/18I/39R. Trump up 1.5 with 2-point party advantage. Inferred ratios, Biden 90-5 D, 47-45I, Trump 90-5 R. Applying to more likely ratios, Biden receives 46.4 to Trump 45.2
46.4
45.2
x
POLLS ABOVE THE RED LINE ARE INCLUDED IN THE AVERAGE
The D+5 sample oversamples Democrats. Adjusting to 34D/29I/33R. According to the data, Biden won D's 83-12, I's 39-34 and Trump won R's 85-10. Applying the ratios to the more likely turnout, Biden receives 42.8 to Trump 42.0.
The D+3 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 92-4, I's 43.5-44.0 and Trump won R's 90-5. Applying the ratios to the more likely turnout, Biden receives 47.8 and Trump 43.8. Biden receives 45.1 and Trump 43.4.
According to the poll data, Biden won D's 95-2, I's 47-37 and Trump won R's 87-10. Applying these voter ratios to the more likely turnout of 34D/30I/33R, Biden receives 49.7 to Trump 40.5.
I looked at all the internals provided behind the pay wall. I can't reveal them, but the numbers look very solid. Rasmussen seems to be catching a trend that seems visible in daily life but doesn't show up much in poll, a large group of Dems tired of the wokeness of the new Dem party.
The D+3 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 94-4, I's 47-44 and Trump won R's 91-5. Applying the ratios to the more likely turnout, Biden receives 47.8 and Trump 43.8. Biden receives 46.7 and Trump 43.8.
I looked at all the internals provided behind the pay wall. I can't reveal them, but the numbers look very solid. Rasmussen seems to be catching a trend that seems visible in daily life but doesn't show up much in poll, a large group of Dems tired of the wokeness of the new Dem party.
The D+3 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 95-3, I's 48-42 and Trump won R's 94-3. Applying the ratios to the more likely turnout, Biden receives 47.8 and Trump 43.8. Biden receives 46.7 and Trump 43.8.
I looked at all the internals provided behind the pay wall. I can't reveal them, but the numbers look very solid. Rasmussen seems to be catching a trend that seems visible in daily life but doesn't show up much in poll, a large group of Dems tired of the wokeness of the new Dem party.
The D+3 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 95-3, I's 49-41 and Trump won R's 95-3. Applying the ratios to the more likely turnout, Biden receives 47.8 and Trump 43.8. Biden receives 47.0 Trump 43.9.
The D+4 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 96-3, I's 47-45 and Trump won R's 95-3. Applying the ratios to the more likely turnout, Biden receives 47.8 and Trump 43.8. Biden receives 46.8 Trump 45.0.
I looked at all the internals provided behind the pay wall. I can't reveal them, but the numbers look very solid. Rasmussen seems to be catching a trend that seems visible in daily life but doesn't show up much in poll, a large group of Dems tired of the wokeness of the new Dem party.
I looked at all the internals provided behind the pay wall. I can't reveal them, but the numbers look very solid. Rasmussen seems to be catching a trend that seems visible in daily life but doesn't show up much in poll, a large group of Dems tired of the wokeness of the new Dem party.
The D+4 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 96-4, I's 49-42 and Trump won R's 93-4. Applying the ratios to the more likely turnout, Biden receives 47.8 and Trump 43.8. Full explanation below.
According to the poll data, Biden won D's 95-2, I's 47-37 and Trump won R's 87-10. Applying these voter ratios to the more likely turnout of 34D/30I/33R, Biden receives 49.7 to Trump 40.5.
The D+5 sample oversamples Democrats. Adjusting to 34D/29I/33R. According to the data, Biden won D's 83-12, I's 39-34 and Trump won R's 85-10. Applying the ratios to the more likely turnout, Biden receives 42.8 to Trump 42.0.
The D+11 sample oversamples Democrats. Will adjust to 34D/30I/33R. According to the data, Biden won D's 85-4, Trump won I's 42-39 and R's 86-5. Applying the ratios to the more likely turnout, Biden receives 42.3 to Trump 42.3.
The D+4 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 95.5-5, I's 49-43 and Trump won R's 91-5. Applying the ratios to the more likely turnout, Biden receives 47.8 and Trump 43.8. Full explanation below.
The D+5 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 92.5-5, I's 47-43 and Trump won R's 91-6. Applying the ratios to the more likely turnout, Biden receives 46.6 and Trump 43.8. Full explanation below.
The D+5 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 92.5-5, I's 47-43 and Trump won R's 91-6. Applying the ratios to the more likely turnout, Biden receives 46.6 and Trump 43.8. Full explanation below.
The D+5 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 90-7, I's 45-41 and Trump won R's 94-4. Applying the ratios to the more likely turnout, Biden receives 44.5 and Trump 44.9. Full explanation below.
The D+5 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 90-7, I's 45-41 and Trump won R's 94-4. Applying the ratios to the more likely turnout, Biden receives 44.5 and Trump 44.9. Full explanation below.
The D+5 sample is wrong and too many I's are included. Adjusting to 34D/28I/33R. This is also effectively an RV sample so an RV filter will apply to the I column. Voter ratios were not provided but can be inferred. Biden had a 5-point partisan advantage and a 9-point lead, suggesting he did 4 points (12%) better across party. Assuming Biden won D's 92-4, I's 54-38 (37-25.3) and Trump won R's 92-6. Applying the voter ratios to the more likely turnout, Biden receives 43.6 to Trump 38.8, or 48.2 to 42.8.
The D+4 sample oversampled Democrats. Will adjust to 37D/24I/36R. According to the poll data, Biden won D's 92-5, I's 47-37 and Trump won R's 89-9. Applying the voter ratios to the more likely turnout, Biden receives 48.6 to Trump 42.8.
The D+4 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 92-6, I's 47-36 and Trump won R's 94-4. Applying the data to the more likely turnout, Biden receives 45.8 to Trump
The D+5 sample oversamples Democrats. Adjusting to 34D/28I/33R. According to the data, Biden won D's 93.5-5, I's 45-38 and Trump won R's 94-4. Applying the voter ratios to the more likely turnout, Biden receives 45.7 to Trump 43.6
The D+5 sample oversamples Democrats. Adjusting to 34D/30I/33R. According to the data, Biden won D's 93-5, I's 46-37 and Trump won R's 94-4. Applying the voter ratios to the more likely turnout, Biden receives 46.7 to Trump 43.8.
1048RV, (Their likely voter sample is not real, just a slightly smaller RV sample). 32D/33I/27R; Leaners 49D/6I/43R.
49
38
x
Political Ref's Take on the above poll
The D+6 sample oversamples Dems and there are too many I's, will adjust to 34D/30I/33R. This is also a registered voter poll so a registered voter screen will apply to the I column. Biden had a 6-point partisan advantage and an 11-point lead, suggesting he did about 5 (15% across three parties) points better across party, assuming Biden won D's 93-3, I's 50-38 (43-35) and Trump won R's 88-9. Applying the new ratios to the more likely turnout, Biden receives 47.5 to Trump 40.5.
Independents are oversampled and Republicans undersampled. Will adjust to 34D/30I/33R. According to the data, Biden won D's 96-4, I's 57-36 and Trump won R's 93-6. Applying the voter ratios to the more likely turnout, Biden receives 51.7 to Trump 42.9.
The D+6 turnout oversamples Democrats. Will assume 42D/13I/41R. Biden has an eleven point lead with a 6-point partisan advantage suggesting he did 5 (15% across 3 parties) points better across party. This is a registered voter poll so a registered voter filter will apply to the I column. Assuming Biden won D's 93-5, I's 54-40 (37-27.3) and Trump won R's 88-10. Applying the voter ratios to the more likely turnout, Biden receives 48.0 to Trump 41.7.
The D+5 turnout oversamples Democrats. Will adjust to 34D/30I/33R. This is a registered voter sample so a registered voter filter will be applied to the I column. According to the data, Biden won D's 87-8, I's 40-32 (38-34) and Trump won R's 86-9. Applying the voter ratios to the more likely turnout, Biden receives 44.00 to Trump 41.3.
The ratings are poor for this poll but it has an 84% race called correctly rate, matching the top two pollsters rated by Silver. Yes, it picks the low-hanging fruit, but it's consistent bias (offers stability) and the large number of jurisdictions it polls are a big plus. This pollster has the most pro-Democrat mean-reverted bias on fivethirtyeight.com, +5.0 more favorable to Dems. (Screenshot, Direct Link). This is also an online only poll, which favor Democrats on average, so an adjustment will be applied. See explanation of the online adjustment.
This is an online only poll, a registered voter poll and fails to report its party distribution and voter ratios. An online poll adjustment will be made. See NYT, Are State Polls Any Better than they Were in 2016.
Will adjust to D+1, or 35D/21I/34R. This is a registered voter poll and a registered voter screen will be applied to the I column (See Voter Turnout Assumptions). According to the data, Biden won D's 86-10, I's 39-32 (38-33) and Trump won R's 82-11. Applying the voter ratios to the more likely turnout, Biden receives 41.8 to Trump 38.3, extrapolate that ratio to the 89% of respondents who chose a candidate and Biden receives 46.4 to Trump 42.6.
The D+5 party distribution oversampled Dems. Adjusting to D+1, or 35D/29I/34R. According to the poll data, Biden won D's 83-10, I's 38-34 and Trump won R's 82-10. Applying the voter ratios to the more likely turnout, Biden receives 43.5 to Trump 41.2.
This is a leaners turnout model. In 2016, the leaners turnout was 51D/48R. I think it will be virtually even this year, or 50D/49R, because Republicans have done better registering voters, have more enthusiasm and many Democrats are voting by mail and will have a larger share of their vote voided. Biden leads by 10 with a partisan advantage of 5, suggesting he did 5 points better across party. Assuming Biden won D and D leans 92-4 and Trump won R and R leans 84-11. Applying these ratios to the more likely voter turnout, Biden receives 51.4 to Trump 43.2. This is also a registered voter poll, so a registered voter filter will be applies, Biden receives 50.1 to Trump 44.5.
1384LV, Party distribution not provided; 1500RV, 37D/35I/27R
LV 51 RV 46
LV 42; RV 38
LV
Political Ref's Take on the above poll
The D+10 turnout oversamples Dems. Using voter ratios for registered voters because likely voters are not provided. According to the poll data, Biden won D's 89-1, Trump won I's 37-32 and R's 92-2. Applying these ratios to the more likely of 33D/32I/32R, Biden receives 40.3 to Trump 41.6.
This poll assumes 78% of registered voters will vote (69% in 2016). This turnout is historically high and contradicts our analysis of the fundamentals of turnout as well as the much more comprehensive Gallup poll, as compared to this poll, released last week. Fox likely included many unlikely voters in its poll, the majority of which probably lean Dem, rendering even our adjustment too favorable for Biden. The D+7 turnout likely oversamples Dems, adjusting to D+1, or 46D/9I/45R. According to the poll data, Biden won D's 93-4, I's 42-28 and Trump won R's 92-6. Applying the voter ratios, Biden receives 49.3 and Trump 45.8.
The D+6 sample oversampled Dems. Will adjust to D+1, or 45D/10I/44R. Voter ratios are not provided but can be inferred. Biden had a 12-point lead with a 6-point partisan advantage, suggesting he did 6 points better (18% across three parties) with Independents and Republicans than Trump did with Independents and Democrats. Assuming Biden won D's 95-1, I's 55-35 and Trump won R's 87-4. Applying these ratios to the more likely turnout, Biden receives 50.0 to Trump 42.2.
This 843LV sample is drawn from 1001 adults, indicating an assumption of a 84.3% turnout, something that has not happened since the Civil War. For this reason, this is essentially a registered voter poll and a registered voter filter will be applied to the entire sample. Adjusted 32/27.3, and Biden receives 48.0 to trump 41.0.
The D+5 sample oversamples Dems and the I+11 sample greatly oversamples Independents. This poll assumes an impossible turnout of 83%. We have not had that kind of turnout since the Civil War. Typically around 60% of Americans vote. This is a tactic some pollsters will use when they find a particularly favorable result for their preferred candidate among Independents. Will apply a registered voter poll screen to the I column. Will adjust to the much more likely D+1, 33D/32I/32R. According to the poll data, Biden won D's 99-1, I's 56-37 (39-24.3 adjusted) and Trump won R's 94-6. Applying the voter ratios to the more likely turnout, Biden receives 47.1 to Trump 38.2.
The D+2 sample slightly oversamples Dems, will adjust to D+1, or 33D/32I/32R. Voter ratios are not provided but can be inferred. Biden had a 2-point turnout advantage and a 1-point deficit, suggesting Trump did 3 points (9%) better with Democrats and Independents than Biden did with Republicans and Independents. Assuming Biden won D's 88-6, Trump won I's 48-42 and R's 91-6. Applying the voter ratios to the more likely turnout, Biden receives 44.4 to Trump 46.5.
1088LV, Party distribution not provided; 1202RV, 40D/30I/27R
LV: 51, RV: 48
LV: 43, RV: 40
LV
Political Ref's Take on the above poll
The D+8 sample oversampled Dems. Will adjust to the more likely D+1, or 40D/20I/39R. According to the poll data, Biden won D's 87-7, I's 51-33 (34-20.3) and Trump won R's 84-3. This is a registered voter result, so a registered voter screen will be applied to the I column, adjusts to 34-20.3. Applying these voter ratios to the more likely turnout, Biden receives, 42.8 to Trump 39.6.
*This is a deceptive poll.* The D+9 sample greatly oversamples Democrats. Will adjust to a D+1 sample, 42D/13I/41R,and apply a registered voter screen to the I column. Voter ratios, or course, are not provided. Biden had a 9-point partisan advantage and a 13-point lead, suggesting he did 4 points (12%) better with Independents and Republicans than Trump did with Independents and Democrats. Assuming Biden won D's 91-3, I's 48-42 and Trump won R's 88-6. Applying a registered voter screen to I's, adjusts to 31-29.3. Applying these voter ratios to the more likely turnout, Biden receives 44.7 to Trump 41.1
The D+4 sample oversamples Democrats. Will adjust to 46D/8I/45R. Voter ratios are not provided but can be inferred. Biden had a 4-point partisan advantage an a 10-point lead, suggesting he did 6 points (18%) better with Independents and Republicans than Trump did with Independents and Democrats. Assuming Biden won D's 94-2, I's 49-40 and Trump won R's 89-6. Applying these ratios to the more likely turnout, Biden receives 49.9 to Trump 44.2.
The D+4 sample oversamples Democrats. Assuming a D+1 voter turnout, or 33D/32I/32R. According to the poll data, Bide won D's 90-8, I's split 42-42 and Trump won R's 92-5. Applying the voter ratios to the more likely turnout, Biden receives 44.7 to Trump 45.5.
1350LV, Party distribution not provided. 1499RV, 37D/34I/29R
50
42
x
Political Ref's Take on the above poll
The D+8 party distribution ovesampled Dems. Adjusting to D+1 or 33D/32I/32R. According to the poll data, Biden won D's 89-3, Trump won I's 35-34 and R's 87-3. Applying the voter ratios to the more likely party distribution, Biden receives 41.2 to Trump 40.0
The D+3 party distribution is a bit high, will adjust to D+1, or 42D/9I/41R. According to the poll data, Biden won D's 91-3, I's 50-25 and Trump won R's 89-6. Applying the voter ratios to the more likely turnout, Biden receives 45.2 to Trump 40.0.
The voter turnout assumption for likely voters was not provided. Will use the registered voter model. The D+6 voter turnout assumption oversamples Dems, will adjust to D+1, or 33D/32I/32R. Voter ratios are not provided but can be inferred. Biden had a 6-point partisan turnout advantage and leads by 6, suggesting that Biden performed evenly with Trump among Independents and cross-party. Assuming each candidate won his party 90-5 and I's split 45-45. Applying these voter ratios to the more likely turnout, Biden receives 45.7 to Trump 44.9.
The D+12 sample oversamples Democrats. Will adjust to the D+3, or 33D/32I/32R. According to the poll data, Biden won D's 87-5, I's split 39-40 and Trump won R's 84-3. Applying these voter ratios to the more likely turnout, Biden receives 44.4 to Trump 41.9.
The D+11 sample oversamples Democrats. Will adjust to the D+3, or 36D/31I/33R. According to the poll data, Biden won D's 84-2, Trump won I's 39-33 and R's 84-6. Applying these voter ratios to the more likely turnout, Biden receives 42.5 to Trump 40.5.
This poll assumes an 86% national turnout among registered voters which will not happen, so many respondents in this poll are not going to vote. Will apply the likely voter turnout to the more likely turnout of 33D/32I/32R. According to the poll data, Biden won D's 91-4, I's 42-25 and Trump won R's 89-6. Biden receives 45.4 to Trump 37.8.
The D+5 sample oversamples Democrats. Will adjust to the more likely voter turnout, 33D/32I/32R. According to the poll data, Biden won D's 82-10, I's 40-32 and Trump won R's 84-7. Applying these updated voter models to the more likely turnout, Biden receives 42.1 and Trump 40.4.
The D+8 sample oversampled Democrats. Will adjust to D+3, or 36D/31I/33R. Voter ratios are not provided by can be inferred. Biden had a +8 partisan advantage and leads by 9, so we can assume he did slightly better with Independents than Trump. Assuming each candidate won his party 90-5 and Biden won I's 48-45% (1 point). Applying the voter ratios to the more likely turnout, Biden receives 48.9 to Trump 45.5
The D+5 sample oversamples Democrats. Will adjust to D+3, or D36/I31/R33. This is effectively a registered voter poll because it assumes 83% of Americans will vote, which has not happened since the Civil War. Typically between 55-65% of Americans voter, as it likely this year. A registered voter filter will be applied to the Independent column. According to the poll data, Biden won D's 84-8, I's 36-33 and Trump won R's 85-17. Adjusting the I column, Biden receives 19-20.3, apply the 19/20.3 ratio to the 69% of I's that chose Biden/Trump, the I columns adjusts to 33-36. Applying the voter ratios to the more likely voter turnout, Biden receives 46.4 to Trump 42.1.
962LV, Party distribution not provided. *Note - Pollster assumes 94% of 1022 registered voters will vote. Also, Biden only leads by 5 among RV, but 6 among LV.
50
44
Political Ref's Take on the above poll
Party distribution is reasonable. No adjustment necessary.
The D+5 sample will be adjusted in the registered voter screen. This is a registered voter poll and includes many unlikely voters, so a registered voter screen will be applied. Adjusted, Biden receives 34 to Trump 30.3. Extrapolating the 34/30.3 ratio to the 94% who chose Biden or Trump, Biden receives 49.7 to Trump 44.3.
The D+12 sample oversamples Democrats. Will adjust to the D+3, or 36D/31I/33R. According to the poll data, Biden won D's 91-2, Trump won I's 43-33 and R's 91-5. Applying these voter ratios to the more likely turnout, Biden receives 44.6 to Trump 44.1.
The likely voter model was not provided, but the likely voter model and registered models show a virtually identical result, so the RV voter model will be assumed. This poll also assumes a 89% turnout among all registered voters, a historic and unlikely turnout. According to the poll data, Biden won D's 94-4, I's 41-37 and Trump won R's 92-5. This is a registered voter model so a registered voter filter will be applies. I's adjusted show Biden 38.8 to Trump 39.2. Applying the voter ratios to the more likely turnout, Biden receives 47.5 to Trump 44.0.
The D+9 sample oversamples Democrats. Will adjust to D+3, or 45D/9I/42R. Voter ratios are not provided but can be inferred. Biden has a 9-point lead and a 9-point partisan turnout advantage, so one can assume they tied with Independents. Assuming each candidate won his party 90-5 and I's split 45-45, Biden receives 48.0 to Trump 45.5.
Party distribution is not provided but based on partial voter ratios, it is clear the party distribution was reasonable. Biden wins 80% of Dems, Trump wins I's by 9 points and 80% of GOP
The D+5 sample oversamples Democrats. Will adjust to D+3, or 36D/31I/33R. This is also a registered voter poll so a registered voter screen will be applied to the I column. According to the poll data, Biden won D's 84-9, I's 37-31 and Trump won R's 82-10. I's adjust to 35.5-32.5. Applying the voter ratios to the more likely voter turnout, Biden receives 44.2 to Trump 40.7.
The party distribution is reasonable, but the overall turnout assumption is not. This Fox poll assumes 91% of all registered voters will turnout. That will not happen. This is not a good likely voter screen. The turnout will likely fall between 65% and 80%.
The D+5 sample oversamples Democrats, but no sample adjustment will be made. This is a registered voter sample, so a registered voter sample will be applied. Biden adjusted receives 30-27.3. Apply the 30/27.3 ratio to the 86% of respondents who chose Biden or Trump, Biden receives 45.0 to Trump 41.0.
823LV, Likely voter party distribution not provided
52
40
Political Ref's Take on the above poll
Party distribution is not provided but voter ratios for LVs are. Biden won D's 93-4, Trump won I's 39-33 and R's 85-8. Will assume a leaners party distribution of 43D/10I/40R. Applying these voter ratios the likely voter turnout, Biden receives 46.5 to Trump 39.6.
Biden has 79% D's, Trump has 80% R's, "nearly tied" with I's. The problem is that so much depends on how many R's Biden gets and D's Trump gets. Without more data, the poll is questionable.
The D+7 sample oversamples Democrats. Will adjust to D+3, 36D/31I/33R. Voter ratios are not provided but can be inferred. Biden has a +7 partisan advantage but leads by 6, so Trump seems to have done about 1 point better with Independents than Biden. Will assume 17-16 (49-46%) with 2 (5%) voting for third parties. Assuming Biden and Trump each won his party 90-5, Biden receives 48.3 to Trump 46.7.
The D+10 sample greatly oversamples Democrats. Average D+ margin is 3.5% in the last nine presidential elections and Democrats face serious challenges to getting out their vote in 2020. Adjusting to D+3, or 36D/31I/33R. According to the poll data, Biden won D's 95-2, I's 47-40 and Trump won R's 91-5. Applying these voter ratios to the more likely voter turnout, Biden receives 50.4 to Trump 43.2. This sample also assumes 98% of registered voters will vote (2433LV/2493RV), which will NOT happen. It will likely be about 70%. It assumes liberals are 34% of the population and conservatives 36%. Suffolk shows 27lib/35con in a openly registered voter poll which squares with Gallup's findings in its all adults survey, 26lib/35con. A likely voter poll should find fewer liberals, not more than a registered or all adult poll. This is a registered voter poll deceptively held out as a likely voter poll. A registered voter filter will be applied. 50.4 43.2 Biden adjusts to 33.4 to Trump 30. Apply the 33.4/30 ratio to the 94% choosing Biden or Trump, and Biden receives 49.5 to Trump 44.5. Yep, we finally got to a realistic number.
The D+15 sample grossly oversampled Dems. Adjusting to D+3, or 36D/31I/33R and applying a registered voter filter to the Independent column. According to the poll data, Biden won D's 91-4, I's split 41-40 and Trump won R's 91-6. I's adjust to 24-27.3, or 37.9-43.1. Applying the voter ratios to the more likely turnout, Biden receives 46.5 to Trump 44.8.
Neither party distribution or voter ratios within those parties are provided. Registered filter will be applied. After registered filter applied, adjusted Biden 32 to Trump 28.3. Applying the 32/28.3 ratio to the 90% of respondents who chose Biden/Trump, Biden receives 47.8 to Trump 42.2.
A registered voter poll will be applied. Adjusted Biden receives 30 to Trump 27.3. Applying 30/27.3 to the 87% of respondents who chose Biden/Trump, Biden receives 45.5 to Trump 41.5.
A registered filter will be applied. Adjusted Biden receives 34 to Trump 30.3. Applying the 34/30.3 to the 94% of respondents who chose Biden or Trump, Biden receives 49.7 to Trump 44.3.
Party distribution is not provided but partial voter ratios are. Biden received 81% of D's, won I's by 3% and Trump received 81% of R's. No adjustment necessary.
Party distribution is not provided but voter ratios are. Biden won D's 90-7, I's split 41 Biden to 37 Trump, and Trump won R's 87-7. Applying the most likely voter turnout model, 36D/31I/33R, Biden receives 47.4 to Trump 42.7.
The D+2 slightly undersampled Democrats. Will adjust to D+3, or 36D/31I/33R. The voter ratios are not provided but can be inferred. Trump had a 3-point lead with a 2-point turnout disadvantage. He likely did better with Democrats and Independents than Biden did with Republicans and Independents. The likely voter screen accounts for this. Will assume Trump did 2 points (6%) better with Dems than Biden did with the GOP and won Independents by 3 (9%) points. Assuming Biden won D's 88-4, Trump won I's 49-40 and R's 94-4, Biden receives 45.4 to Trump 47.7.
The D+17 sample greatly oversamples Democrats. It looks like YouGov uses an adjusted 41D/30I/29R sample, but it reports a D+17. Will adjust to the much more likely D+3, or 36D/31I/33R. According to the poll data, Biden won D's 88-5, Trump won I's 44-34 and R's 88-6. Applying the voter ratios to the more likely voter sample, Biden receives
The party distribution looks good. This is a registered voter poll so a registered voter screen will be applied. Biden loses 17 and Trump loses 12.7, results in 30-27.3. Applying this 30/27.3 ratio to the 87% of respondents who chose a candidate, Biden receives 45.5 to Trump 41.5.
The D+5 sample oversampled Democrats. Will adjust to D+3, or 36D/31I/33R. This is also a registered voter poll. A registered voter screen will be applied to the Independent column. According to the poll data, Biden won D's 87-6, I's 42-29 and Trump won R's 83-7. Applying the registered voter screen to the Independent column, Biden loses 5.0 and Trump loses 3.7, resulting in Independents voting 7.2-4.7 for Biden with the rest staying home or voting third party. So Biden win I's 25-16.2%. Apply the voter ratios mentioned above to the more likely voter turnout, Biden receives 41.4 to Trump 34.6. Applying the 41.4/34.6 ratio to the 85% who chose Trump or Biden, Biden leads 46.3 to 38.7.
Party distribution is not provided but partial voter ratios are. Biden and Trump earned even support (76% & 77%) among their parties and even support among Independents.
The D+14 sample far oversamples Democrats. Will adjust to the much more likely 36D/31I/33R voter turnout assumption. Will also apply a registered voter screen to the Independent column as many Independents will not vote who are included in this sample. According to the poll data, Biden won D's 91-4, I's 46-35 and Trump won R's 89-4. Applying the registered voter screen to the independent column, Biden drops to 29 and Trump to 22.3. Applying the voter ratios to the more likely voter turnout, Biden receives 43.1 to Trump 37.7.
Biden had an 8-point lead with an 8-point party turnout advantage. The candidates, therefore, performed evenly with Independents. Adjusting to a much more likely D+3 turnout advantage, or 36D/31I/33R. Assuming each candidate wins his party 90-5 and splits Independents, Biden receives 48.1 to Trump 45.5.
The D+5 and I+11 samples do not approach anything resembling reality. This is not a sample that anyone believes, but the goal is to get a headline with a big Biden lead so mission accomplished. Voter ratios are not provided and the voter turnout assumption is so flawed that any adjustment is impossible because there is simply not enough data.
This poll contains far too many Independents. Independents have not outnumbered either Dems or the GOP in the last nine presidential elections. It's unclear what legitimate reason pollsters have in inflating the I column so high, except that in registered voter polls in tends to bump the Democrat considerably because many Dem leaning Independents tend not to vote on election day, but can be counted in polls. Will adjust to the much more likely 36D/31I/33R. Biden had a 4-point turnout advantage and leads by 4, suggesting the two candidates broke even among Independents. With a 3-point advantage in turnout, therefore, Biden would be leading by 3. Adjusted result, Biden 49.0, Trump 46.0.
The D+10 sample oversamples Democrats. Will adjust to the much more likely 36D/31I/33R party distribution. According to the poll data, Biden won D's 94-4, Trump won I's 45-40 and R's 91-6. Applying these voter ratios to the more likely turnout, Biden receives 48.2 to Trump 45.4.
The party distribution looks reasonable, but this is a registered voter poll so I will apply the registered voter screen. Subtract 17 from Biden and 12.7 from Trump to reach 33 to 28.3. Apply the 33/28.3 ratio to the 91% of voters who chose Biden or Trump. Biden receives 48.9 to Trump 42.0.
The voter distribution seems reasonable, but this is a registered voter poll so I will apply a registered voter screen. Biden loses 17 and Trump loses 12.7 to reach 36 to 29.3. Apply the 36/29.3 ratio to the 95% of respondents who chose Biden or Trump and Biden receives 52.4 to Trump 42.6.
The party distribution looks reasonable. According to the poll data, Biden won D's 91-4, Trump won I's 30-28 and R's 86-8. Applying the voter ratios to the reported party distribution, Biden receives 48.4 to Trump 42.1.
The D+13 sample oversamples Democrats. Will adjust to the more likely 45D/9I/42R voter turnout assumption to match 2016 in the D+3 component. According the poll data, Biden won D's 86-6, I's 32-27 and Trump won R's 84-7. Applying these voter ratios to the much more likely voter turnout, Biden receives 44.5 to Trump 40.4.
The D+5 slightly oversamples Democrats. Adjusting to 36D/31I/33R. Voter ratios are not provided but can be inferred. Biden had a +5 voter turnout advantage but trailed by 4, suggesting Trump did slightly better with Independents than Biden. Assuming Trump won I's 14-13, or 48-45% with the rest voting third party, and that each candidate won his party 90-5, Biden receives 48.0 to Trump 46.4.
This is a registered voter poll that includes many voters who will not vote. A registered voter screen will be applied. Biden loses 17 and Trump 12.7 to reach 34 to 28.3. Apply the 34/28.3 ratio to the 92% of respondents who chose Biden or Trump and Biden receives 50.2 to Trump 41.8.
The D+5 sample slightly oversampled Democrats. Adjusting to 36D/31I/33R. According to the poll data, Biden won D's 87-5, I's 34-22 and Trump won R's 87-9. Applying these voter ratios to the more likely voter sample, Biden receives 44.8 to Trump 37.3. Applying now the registered voter screen, Biden drops to 27.8 to Trump 24.6 and apply the 27.8/24.6 ratio to the 82% of respondents who chose Biden or Trump, and Biden receives 43.5 to Trump 38.4.
The D+10 sample oversamples Democrats. Will adjust to the more likely 36D/31I/33R voter turnout. According to the poll data, Biden won D's 88-9, I's 48-42 and Trump won R's 88-8. Applying these voter ratios to the more likely turnout, Biden receives 49.2 to Trump 45.3.
The party distribution looks reasonable. This is a registered voter poll, however, so a registered voter screen will be applied. Biden loses 17 and Trump loses 12.7 to reach 36 to 32.3. Apply the 36/32.3 ratio to the 98% of the respondent who chose Biden or Trump and Biden receives 51.7 to Trump 46.3.
The D+5 sample oversamples Democrats. Will adjust to the more likely D+3, or 36D/31I/33R sample. According to the poll data, Biden won D's 82-10, Trump won I's 35-33 and R's 81-10. Applying these ratios to the more likely turnout, Biden receives 43.1 to Trump 41.2.
According to the poll data, Biden won D's 86-7, I's 39-32 and Trump won R's 83-10. Applying these ratios to the most likely voter turnout of 36D/31I/33R, Biden receives 46.4 to Trump 39.8.
The D+7 sample oversampled Democrats. Will adjust to the more likely D+3, or D36/I31/R33. According to the poll data, Biden won D's 90-4, I's split 41-39 for Biden and Trump won R's 89-5. Applying the ratios to the more likely turnout, Biden receives 46.8 to Trump 42.9.
The D+19 sample grossly oversamples Democrats. Will adjust to the much more likely 36D/31I/33R voter turnout model. According to the poll data, Biden won D's 88-5, Trump won I's 45-36 and R's 88-5. Reallocating these voter ratios to the more likely voter turnout model, Biden receives 44.5 to Trump 44.8.
Party distribution was not provided by voter ratios within the parties was. Biden won D's 90-6, I's 47-28 and Trump wins R's 86-7. Applying those ratios to the most likely voter turnout as we see it, 36D/31I/33R, Biden receives 49.3 to Trump 39.2.
The D+13 sample oversamples Democrats. The I+12 sample (more I's than R's) oversamples Independents. Adjusting to the much more likely voter turnout of D+3 and I-2, or 36D/31I/33R. According to the poll data, Biden won D's 89-4, Trump won I's 40-38 and R's 88-6. Applying these voter ratios to the more likely vote turnout, Biden receives 45.8 to Trump 42.9. Full explanation below.
The D+8 sample oversamples Democrats. Adjusting to the much more likely D+3 we saw in 2016, or 45D/8I/42R. According to the poll data, Biden won D's 86-5, I's 39-23 and Trump won R's 83-7. Applying these voter ratios to the more likely turnout, Biden receives 44.8 to Trump 39.0. Full explanation below.
The D+6.5 over R oversamples Democrats and the I+7.8 oversamples Independents. Adjusting to the much more likely 36D/31I/33R turnout. According to the poll data, Biden wins D's 80-6, I's 24-21 and Trump wins R's 76-8. Applying these ratios to the more likely voter turnout, Biden receives 38.8 to Trump 33.8. Applying the 38.8/33.8 ratio to the 88% of respondents who chose Biden or Trump, Biden receives 47.0 to Trump 41.0. Full explanation below.
Will assume a voter turnout of 36D/31I/33R. Voter ratios are not provided but can be inferred. Biden is up 9 but had a 5-point turnout advantage, suggesting he did 4 points better overall with Republicans and Independents than Trump did with Democrats and Independents. Assuming the difference all came with Independents to make the calculation simpler, or that Biden won I's 17.5-13.5, or 56%-44%. Assuming also that each candidate won his party 90-5. Biden receives 51.4 to Trump 45.1.
Party distribution is not provided by voter ratios withing the parties are. Biden won D's 87-5, I's 35-25 and Trump won R's 84-7. Applying these voter ratios to a national voter turnout of 36D/31I/33R, Biden receives 44.5 to Trump 37.3.
The D+9 sample oversamples Democrats. Adjusting to the much more likely D+3 sample, 36D/31I/33R. According to the poll data, Biden won D's 93-3, Trump won I's 43-40 and Trump won R's 88-8. Reallocating these voter ratios to the more likely turnout, Biden receives 48.5 to Trump 43.5.
Data for Progress does not provide its voter turnout assumption or voter ratios within the parties. This is, however, a registered voter poll and a registered voter poll can be applied. See the Voter Turnout Assumptions section above to see how the registered voter poll filter applies. Biden loses 17 and Trump loses 12.7 to 32.8 and 31.2 respectively. Apply the 32.8/31.2 ratio to the 93.7% of voters who chose Trump or Biden in this poll. Final adjusted result is Biden 48.0 to Trump 45.7.
Party distribution is not provided, but voter ratios are. Biden won 74% of D's, I's 47-36 and Trump won 81% of R's. Assuming each candidate won 5% of the other party's vote, after applying the voter ratios to the likely voter model from 2016, 36D/31I/33R, Biden receives 42.9 to Trump 39.7. Applying the 42.9/39.7 ratio to the 92% of respondents who chose either Biden or Trump, Biden receives 47.8 to Trump 44.2.
The D+8 sample oversamples Democrats. Will adjust to the much more likely D+3 2016 turnout model, 42D/10I/39R. According to the poll data, Biden won D's 83-6, I's 29-23 and Trump won R's 80-8. Reallocating these voter ratios to the more likely voter turnout model, Biden receives 40.9 to Trump 36.0.
The D+5 sample oversamples Democrats. Will adjust to the more likely D+3, 36D/31I/33R voter turnout from 2016. According to the poll data, Biden wins D's 84-7, I's 38-31 and Trump wins R's 81-9. Applying these voter ratios to the more likely voter turnout, Biden receives 45.0 to Trump 38.9.
Neither party distribution or voter ratios within parties is provided. This is, however, a registered voter poll so a registered voter screen can be applied. See "How registered voter polls can be adjusted" above for an explanation. We subtract 17 from Biden and 12.7 from Trump to get 30 Biden to 27.3 Trump. Apply the 30/27.3 ratio to the 87% of the voters who chose either Biden or Trump to receive Biden 45.5 to Trump 41.5.
The D+12 sample oversamples Democrats. Adjusting to the more likely D+3 from 2016, 36D/31I/33R. According to the poll data, Biden won D's 91-4, I's split 37 for Biden to 38 for Trump and Trump won R's 88-7. Reallocating these voter ratios to the more likely 36D/31I/33R voter turnout, Biden receives 46.5 to Trump 42.3.
The Fox News poll of registered voters fails to apply its own screen for likely voters. Among those included in its poll results are 7% who report they are only somewhat likely to vote and 2% who say they are not at all likely to vote. We know from past elections that these unlikely voters lean Democratic. Those voters should be screened out of the results, but they fail to do that so we are stuck with these unlikely results. Because of this, I can only apply the relatively imprecise registered voter poll adjustment. For an explanation of this adjustment, see this page. Subtract 17 from Biden and 12.7 from Trump to reach 32 Biden to 28.3 from Trump. Apply the 32/28.3 ratio to the 90% of respondents choosing Biden or Trump to reach Biden 47.8 to Trump 42.2.
673LV, Party distribution among likely voters not provided. Among registered voters, it was 30D/39I/24R.
54
44
Political Ref's Take on the above poll
Neither the party distribution or the voter ratios within parties were provided for the likely voter sample. Adjustment is not possible. Excluded from our averages.
Adjustment is not possible. Excluded from our averages.
Neither party distribution or voter ratios within those parties are provided. Because of this, I can only apply the relatively imprecise registered voter poll adjustment. For an explanation of this adjustment, see this page. Subtract 17 from Biden and 12.7 from Trump to reach 32 Biden to 28.3 from Trump. Apply the 32/28.3 ratio to the 90% of respondents choosing Biden or Trump to reach Biden 47.8 to Trump 42.2.
The D+6 sample oversamples Democrats. Will adjust to a more likely D+3, 44D/11I/41R. According to the poll data, Biden won D's 85-3, I's split at 32 Biden 28 Trump, and Trump won R's 82-7. Reallocating those voter ratios to the more likely voter turnout, Biden receives 43.8 to Trump 38.0. Applying that 43.8/38.0 ratio to the 84% of respondents who chose Biden/Trump, Biden receives 45.0 to Trump 39.0
The D+14.9 sample oversamples Democrats. Adjusting to the much more likely 36D/31I/33R voter turnout assumption. According to the poll data, Biden won D's 88-4, I's split at 39-39 and Trump won R's 87-4. Reallocating the voter ratios to the more likely voter turnout model, Biden receives 45.1 to Trump 42.2.
The D+10 sample oversamples Democrats. Adjusting to the much more likely D+3, 36D/31I/33R. According to the data, Biden won D's 91-5, I's 51-34 and Trump won R's 84-9. After reallocating the voter ratios to the more likely voter turnout model, Biden receives 51.5 to Trump 40.1. It is important to remember that this is a registered voter poll, which almost always include more Democratic leaning respondents who will not actually vote than Republican respondents who will not actually vote.
The D+7 sample oversamples Democrats. Adjusting to the much more likely D+3, 41D/11I/38R. Voter ratios are not provided, so only the slightly clumsy less accurate adjustment can be made. Biden had a +7 voter turnout assumption advantage but leads by 11, suggesting he is performing strongly with Independents and Republicans. Because of Biden's strength with Independents and Republicans, shifting 2 points out of the D Column and 2 points into the R column will not yield a 4 point gain for Trump, but more like a 3 point gain. Adjusted result, Biden receives 49.5 to Trump 41.5.
Party distribution is not provided, but voter ratios are. Biden won 76% of D's, I's 44-38 and Trump won 79% of R's. Assuming each candidate won 5% of the other party's vote, after applying the voter ratios to the likely voter model from 2016, 36D/31I/33R, Biden receives 42.7 to Trump 39.7. Applying the 42.7/39.7 ratio to the 91% of respondents who chose either Biden or Trump, Biden receives 47.2 to Trump 43.8. It looks like this is the approximate voter sample used in this poll.
The D+5 sample oversamples Democrats, an assumption backed up by the internals. This is one of the only polls showing its pool of likely voters favoring Biden on the economy and by 4 points. Will adjust to a D+3 voter turnout assumption, 46D/11I/43R. Biden had a +5 advantage in the turnout assumption but leads by 10. Trump gains only modestly in the adjustment, Biden 50, Trump 42.
Neither party distribution or voter ratios within parties is provided. This is, however, a registered voter poll so a registered voter screen can be applied. See "How registered voter polls can be adjusted" above for an explanation. We subtract 17 from Biden and 12.7 from Trump to get 31 Biden to 26.3 Trump. Apply the 31/26.3 ratio to the 87% of the voters who chose either Biden or Trump to receive Biden 47.1 to Trump 39.9.
The D+13 sample oversamples Democrats. Adjusting to the much more likely 36D/31I/33R. According to the poll data, Biden won D's 91-4, I's 44-35 and Trump won R's 88-6. Reallocating these voter ratios to the more likely vote turnout, Biden receives 48.4 to Trump 41.3
The party distribution is not provided, but the voter ratios are. Biden won D's 88-6, I's 38-30 and Trump won R's 87-6. Reallocating these ratios to the 2016 voter turnout, Biden receives 45.4 to Trump 40.2.
Neither party distribution or voter ratios within parties is provided. This is, however, a registered voter poll so a registered voter screen can be applied. See "How registered voter polls can be adjusted" above for an explanation. We subtract 17 from Biden and 12.7 from Trump to get 31 Biden to 26.3 Trump. Apply the 31/26.3 ratio to the 87% of the voters who chose either Biden or Trump to receive Biden 47.1 to Trump 39.9.
The voter ratios as applied to the reported turnout do not seem to add up to the reported result. According to the poll data, Biden won D's 80-6, I's split 29-29 and Trump won R's 80-6. Applying these voter ratios to the reported voter turnout of 44.3D/9.7I/42.1R, Biden receives 40.2 to Trump 39.2. I'm not sure what happened, but their reported numbers do not add up.
The D+18.2 sample oversamples Democrats. Adjusting to the more likely 36D/31I/33R voter turnout model. According to the poll data, Biden won D's 89-5, I's split with 39 Biden to 40 for Trump, and Trump won R's 91-2. Reallocating the voter ratios to the more likely voter turnout, Biden receives 44.8 to Trump 44.2.
Party distribution is not provided, but voter ratios within parties were. According to the poll data, Biden won D's 90-7, I's 50-36 and Trump won R's 82-11. Reallocating to the 2016 voter turnout of 36D/31I/33R, Biden receives 51.5 to Trump 40.8.
The D+8 sample oversamples Democrats. Adjusting to the more likely 36D/31I/33R voter distribution. Voter ratios within parties are not provided, but can be inferred. Biden had a +8 voter sample and leads by 11. This suggests he won Independents by around 3, or 12.5-9.5 or 57-43%. Assuming each candidate won his party 90-5, which is unlikely but we are assuming Independents represent the full shift so it will work for calculation purposes, and Biden won Independent 57-43%, Biden receives 51.7 to 44.8 Trump.
The D+7.4 sample oversamples Democrats. Adjusting to the more likely 36D/31I/33R voter turnout. According to the poll data, Biden won D's 86-5, Independents were split at 38 Biden to 35 Trump, and Trump won R's 87-6. Reallocating these voter ratios to the more likely turnout, Biden receives 44.7 to Trump 41.4. Applying the 44.7/41.4 ratio to the 85% of the respondents who chose Biden/Trump, the result is Biden 44.1 to 40.9.
The sample clearly includes too many Independents. Biden wins Independents in this poll 53-36%. Other poll analysts have noted the problematic nature of including so many Independents such as this pollster writing in The Hill. We also saw this with the NYT/Sienna poll and the CNN poll, all three showing huge leads for Biden derived from a very lopsided lead for Biden among Independents. These polls are out of step with other pollsters, like YouGov and Optimus which both show a very close race among Independents. When digging into Monmouth's internals the problem with including so many Independents become evident. When asked how likely respondents were to vote, 88% of Republicans and 90% of Democrats answered certain to vote, while only 78% of Independents did. Only 3% of Republicans and 4% of Democrats said they were not sure or unlikely to vote while 7% of Independents did. We already know that Democratic leaning Independents show up less than Republican leaning Independents. Despite these glaring signs that many Independents Monmouth is polling will not vote, all the Independent respondents are included. I will include this poll in the pollster averages but not in our average until they start screening for likely voters. These wildly inaccurate numbers throw off poll averages. If you doubt the wisdom of excluding this poll and others like it from my average, check to see which line chart more closely resembles the final result on election day, the pollster chart or the one showing our averages. The comparison should be quite telling. If you want to know where the race really stands, come here.
The D+19.6 sample oversamples Democrats. Will adjust to the much more likely 36D/31I/33R voter turnout model. According to the poll data, Biden won D's 88-5, I's split at 40 Trump to 38 Biden and Trump won R's 91-4. Reallocating the voter ratios to the more likely turnout model, Biden receives 44.8 to Trump 44.2.
The D+10 sample oversampled Democrats. Adjusting to the more likely 36D/31I/33R voter turnout. According to the poll data, Biden won D's 83-5, I's 32-24 and Trump won R's 85-5. Reallocating the voter ratios to the more likely party distribution turnout, Biden receives 41.5 to Trump 37.3. Applying the 41.5/37.3 ratio to the 84% of respondents who chose Biden/Trump, Biden receives 44.2 to Trump 39.8.
According to the poll data, Biden won D's 91-4, I's 40-34 and Trump won R's 89-8. Applying these ratios to the likely turnout of 36D/31I/33R, Biden receives 47.8 to Trump 41.3. It is important to note that this is a registered voter poll and includes many voters who will not vote that likely lean Democratic. The margin is likely closer than even this adjusted result.
Neither party distribution or voter ratios within parties is provided. This is, however, a registered voter poll so a registered voter screen can be applied. See "How registered voter polls can be adjusted" above for an explanation. We subtract 17 from Biden and 12.7 from Trump to get 30 Biden to 27.3 Trump. Apply the 30/27.3 ratio to the 87% of the voters who chose either Biden or Trump to receive Biden 45.5 to Trump 41.5.
The D+6 sample oversamples Democrats. Adjusting to the more likely 36D/31I/33R. Voter ratios are not provided but can be inferred. Biden had a +6 advantage in the turnout model and won by 8, so we can assume he won Independents by 2, 18-16, or 53-47%. Extrapolating these ratios to the more likely voter turnout, Biden receives 50.5 to Trump 46.1. Applying the 50.5/46.1 ratio to the 90% of the voters who chose Trump or Biden, the final adjusted result is Biden 47.0 to Trump 42.9.
The D+5.8 sample oversamples Democrats. Will adjust to the more likely 36D/31I/33R voter turnout model. Voter ratios are not provided but we can infer them. Biden has a 6-point advantage in turnout and wins by 9. For that reason, we can assume Biden won Independents 16.6-13, or 57% to 43%. Reallocating the voter ratios to the new turnout model, Biden receives an initial result of 51.7 to Trump 44.8. Applying that 51.7/44.8 ratio to the 83% of the vote that selected Trump or Biden, the final result is Biden 44.5 to Trump 38.5. It is important to note that this is a registered voter poll and includes many voters who will not vote that likely lean Democratic. The margin is likely closer than even this adjusted result.
The D+5 sample oversamples Democrats. Will adjust to the more likely 36R/31I/33R voter turnout. According to the poll data, Biden won D's 90-5, I's split 36 Biden to 33 Trump, and Trump won R's 86-8. Reallocating the voter ratios to the more likely turnout of 36D/31I/33R, results in Biden 46.2 to Trump 40.4.
The D+6.5 sample oversamples Democrats. Will adjust to the more likely 36D/31I/33R voter turnout. According to the poll data, Biden won D's 83.5-3.3, I's 25.6-20.3 and Trump won R's 81.5-6. Reallocating the vote ratios to the more likely turnout of 36D/31I/33R results in Biden receiving 40.0 to Trump 34.4.
The D+5 sample oversamples Democrats. Will adjust to 2016 voter turnout of 36D/31I/33R. According to the poll data, Biden won D's 82-8, I's 36-30 and Trump won R's 82-8. Reallocating the voter ratios to the more likely voter turnout results in Biden 43.3 to Trump 39.2.
The D+8.9 sample clearly oversamples Democrats. Will adjust to the much more likely 36D/31I/33R voter turnout we saw in 2016. According to the poll data, Biden won D's 89-3, I's split with Biden 39 to Trump 37, and Trump won R's 82-8. Reallocating these ratios to the more likely turnout results in Biden 46.8 to Trump 40.2. Keep in mind that this is a poll of registered voters and Democratic leaning registered voters typically fail to vote in as high a turnout percentage of Republican leaning registered voters. For this reason, the margin is likely closer than this number suggests, but because the pollster chose not to screen out unlikely votes we are stuck with this unlikely result. The margin, if an election were held today, would likely be closer than this adjusted result.
The D+7 sample oversamples Democrats. Adjusting to the much more likely voter turnout of 36D/31I/33R. According to the poll data, Biden wins D's 95-3, I's split at Biden 47 and Trump 44, and R's vote Trump 91-7. Reallocating these ratios to the more likely voter turnout results in Biden 50.9 to Trump 44.8. This is also a registered voter poll, which means many of the respondents in this poll will not vote, between twenty to forty percent. Keep in mind that this is a poll of registered voters and Democratic leaning registered voters typically fail to vote in as high a turnout percentage of Republican leaning registered voters. For this reason, the margin is likely closer than this number suggests, but because the pollster chose not to screen out unlikely votes we are stuck with this unlikely result.
The D+10.4 sample clearly oversamples Democrats. Adjusting to the much more likely voter turnout of 36D/31I/33R. According to the poll data, Biden wins D's 85-6, I's are tied at 24-24 and Trump wins R's 82-8. Reallocating the votes in the more likely turnout model results in Biden 40.7 to Trump 36.7. Applying the 40.7/36.7 ratio to the 84% of the vote that chose Biden or Trump in the poll results in Biden 44.2 to Trump 39.8.
The +11.8 D sample oversamples Democrats. Will adjust to the much more likely voter turnout of 36D/31I/33R. According to the poll data, Biden won D's 89-6, I's 43-35 and Trump won R's 89-7. Allocating this Biden/Trump ratio of each party to the 90% of the electorate choosing one of the two candidates results in Biden receiving 47.7 to Trump 42.4. Remember that this poll only includes registered voters, so no doubt that many of the included respondents in this poll will not vote. Among those non-voters, more would likely vote for the Democrat than the Republican.
The D+9 sample oversamples Democrats. Will adjust to the more likely party distribution of 36D/31I/33R. According to the poll data, Biden won D's 90-5, I's 47-29 and Trump won R's 90-5. Allocating these ratios to the more likely 36D/31I/33R turnout, Biden receives 48.6 to Trump 40.5. There is a strong chance that Sienna failed to pick up some of the Trump vote among Independents as noted by other poll analysts. The results among Independents are far out of step with other pollsters like YouGov showing a dead heat in the group. Furthermore, while I have adjusted the party distribution, even my adjustment includes responses of a substantial number of people who will not vote. Because, almost as a rule, more poll respondents who choose the Democrat will fail to actually vote than those who choose Republican candidates. Because this is true the ratios are incorrect. If a likely voter filter were applied it is likely the Democratic share of each ratio would reduce in size. Consider this poll for example, where Biden won D's 90-5, I's 47-29 and Trump won R's 90-5. It's likely that these ratios might look more like the following once a likely voter filter were applied, Biden wins D's 88-7, I's 43-33 and Trump wins R's 90-4. If these were the ratios, Biden would have received 46.3% and Trump 42.5, a 4.1% margin instead of a 14% margin. Hillary won the popular vote by 2.1% yet Trump still won. If Trump is only down 4 instead of 14, he is already in striking distance of winning. So we see that this poll includes many respondents who will not vote, which almost always includes more Democratic leaning voters than Republican voters, but because the NYT did not screen for likely voters we are stuck with the less accurate and pro-Democratic registered voter result. To further demonstrate, 13% of Independents, 5% of Republicans and 8% of Democrats included in this poll stated they were only "somewhat likely," "not very likely" or "not at all likely" to vote, yet their responses were all included in this poll. These highly unlikely voters represent low hanging fruit for a pollster seeking to filter out unlikely voters, yet the New York Times includes their responses. One might reasonably conclude that the inclusion of these responses in a political poll undermines the credibility of its results and I do. I am excluding this poll from my average.
Although party distribution and Biden/Trump ratios of the vote within each party are not provided, a simple registered to likely voter adjustment can be made, see "How Registered Vote Polls Can be Adjusted" above. Subtracting 17 from Biden and 12.7 from Trump, we end up with Biden 30 to Trump 26.3. Allocating the 30/26.3 ratio to the 86% of the vote composed of Biden/Trump voters in this poll results in Biden
The D+10 sample oversamples Democrats. A D+3 is more likely. This is a rare instance where an adjustment is possible without the ratios of how each party is voting because it's obvious from the results. Biden had a 10 point party advantage in the sample and is leading by almost 10, so the Independent vote likely split evenly. Allocating the party distribution to 36D/31I/33R, Dems lose 6.5 from the 7 deducted from the D column, and this is the rare situation where we are safe in assuming everything else remains as is. The D/R ratio isn't provided, but it looks like Independents split fairly evenly here considering it was a +10 D sample and +11 D final result. Adjusted result, Biden 45.5, Trump 43.0.
The D+6.5 sample oversamples Democrats. Will adjust to the more likely 36D/31I/33R turnout. According to the poll data, Biden won D's 80.3-7.2, I's were split at 30.3 Biden to 29.9 Trump, and Trump won R's 81.9-9.1. Allocating these Biden/Trump party ratios to the likely 36D/31I/33R turnout results in Biden 41.3 to Trump 38.9. The remainder in this poll said they would not vote. Allocating the 41.3/38.9 ratio to the 94% of the electorate voting Biden/Trump in this poll results in Biden 48.4 to Trump 45.6. This is a likely voter poll so no registered voter screen is necessary.
The D+8 sample oversamples Democrats. Will adjust to the 2016 party distribution of 36D/31I/33R. According to the poll data, Biden won D's 93-4, I's 43-40 and Trump won R's 92-7. Apply the Biden/Trump ratios of each party vote to the party distribution of 36D/31I/33R results in Biden 49.1 to Trump 44.2.
The D+6 sample oversamples Democrats as seen in the party distribution of 47D/13I/41R. It should be 44D/13I/41R if we are using 2016 as a guide. Because the pollster withholds the vote distribution within each party, we are left to infer the ratio of Biden/Trump voters within each party. I will be assuming a basic turnout model of 36D/31I/33R, after deducting 3 points from the D column, we will assume that 44D/16I/41R contains this basic partisan distribution with leaner Independents making up the rest of each party's makeup. So while 36D is actual Democrats, 44D adds 8 points of Dem leaning I's. Same for GOP, 41R contains 8 points of R leaning I's. Biden loses 2.8 from the 3 deducted from the D column and Trump loses 0.2. Biden gains 1.6 and Trump gains 1.4 from the 3 added to the I column. Because this is a poll of registered voters, an additional screen must be added (See REGISTERED VOTER SCREEN above for an explanation). Biden loses 17 and Trump loses 12.7, resulting in 31.6 Biden to 26.5 Trump. Extrapolating that ratio out to the 88% of the vote composed by D's and R's, the final adjustment is Biden 47.9 to Trump 40.1. The bottom line, this poll includes far too many respondents who will not vote, which casts serious doubt on its credibility. Excluded from our averages.
The D+9 sample is too high of an estimation of Dem turnout. Will reallocate the partisan vote to the 2016 party distribution of 36D/31I/33R. According to this poll's data, D's voted 94-3 for Biden, R's voted 90-6 for Trump and I's voted 41-39 for Trump. Allocating this partisan vote to the 2016 party distribution results in Biden 47.9 to Trump 43.5.
The poll's party distribution isn't provided, but the vote of each partisan group is. I will apply that vote to a 36D/31I/33R party distribution like we saw in 2016 and close to 2018. D's voted 89-5 for Biden, R's voted 82-9 for Trump, I's voted 39-26 for Biden. Initial adjusted result is 47.1 Biden to 36.9 Trump.
This polls uses a D+6.5 party distribution, too high on the Dem side. Adjusting to a D+3 like 2016 and 2018, using a party distribution of 36D/31I/33R. According to the data, D's voted 78.6-7.5 for Biden, R's voted 77.8-9.4 for Trump, I's voted 34.1-23.9 for Biden. Remainder of each group would not vote. Initial adjusted result, Biden 42.0 to Trump 35.8. Extrapolating to the 94.5% of the vote the poll is estimating, final result is 51.0 Biden to 43.5 Trump.
This is a D+8.1 party distribution, too high on the Democratic side. The poll data showed Biden wins Dems 90-4 and Indys 45-37. Adjusting to the D+3 advantage we saw in 2016, Biden loses 4.6 from the D column and Trump loses 0.2. Biden gains 2.3 from the 5.1 added to the I column and Trump gains 1.9. Adjusted result, Biden 46.2, Trump 41.7.
I will adjust to a 36D/31I/33R party distribution. The poll data shows Biden won Dems 90-5, Indys 42-39 and Trump won R's 88-6. Applying the Biden/Trump ratio to each party in the most likely 2016 party distribution results in Biden 47.4 to Trump 42.9.
The D+5.7 sample is too high. Adjusting to the 2016 party distribution of 36D/31I/33R. According to the poll data, Biden won D's 82-5, I's 34-32 and Trump won R's 78-8. Applying the Biden/Trump ratio to each party in the most likely 2016 party distribution results in Biden 42.7 and Trump 37.5. This poll has a lot of people who claim they will not vote, so to get a more realistic outcome, we eliminate those non-voters by applying this 42.7/37.5 Biden/Trump ratio to the 84% of the voters who chose one of the two major candidates. This results in Biden 44.7 to Trump 39.3.
The voter distribution results in a D+6.5. Adjusting to a more likely D+3.0. New distribution is 31.5D/39.8I/28.5R. Biden loses 3.5 but gains 1.8 from the 3.5 added to the I column. Trump gains 1.7 from the 3.5 added to the I column. Adjusted result, Biden 51.4, Trump 43.6.
The D+12 sample oversamples Democrats. Adjusting to a much more likely D+3, with 44D/13I/41R. Biden holds a seven-point lead with a twelve point advantage in the voter model. We do not have the Biden/Trump voter ratios but we can infer them. Trump had to overcome a twelve-point disadvantage in the voter model yet he only trails by seven. We can assume he wins Independents and/or fairs well with Democrat voters. For purposes of analysis we will assume he did average with Democrats and that he won Independents by 5 points, 9.5-3.5, or 73-27. In reality this wouldn't happen as some of that would come from Democrats instead of Independents, but for analysis we will assume it does, and that doesn't harm our analysis. Assuming Biden wins D's 90-5, Trump wins I's 73-27 and Trump wins R's 90-5, results in Biden 46.4 to Trump 48.7.
This poll heavily oversampled Democrats, +10.4%. Trump wins Indy's 42-36, more Democrats than Biden wins Republicans (8-6), and more Republicans than Biden wins Democrats (88-86). Trump should be ahead if they didn't have a large oversample of Democrats. D+10.4 is definitely not going to happen. D+5 is more reasonable, although still better for Dems than 2016 when they won the popular vote. So the new distribution is 33.1D/32.9I/28.1R. Biden loses 5.4 from the Dem reduction but gains back 2.5 from the 5.4 added to the I column. Trump gains 2.9 from the 5.4 added to the I column because he wins Indys in this survey. Adjusted result, Biden 43.1, Trump 44.9.
The party distribution of this poll is 40D/32.4I/27.6R. Trump wins more Dems than Biden wins Rs, 6% to 4%. Trump wins more of his own party than Biden wins of his, 91% to 88%. Biden does win I's, but only barely, 39-37. Yet Biden leads by 5 points? You can only explain this with the large number of Democrats polled over Republicans, a D+12.4% sample. New distribution is 32.6D/39.8I/27.6R. I'm assuming D+5, Biden loses 7.4 from the Dem reduction but gains 3.9 from the 7.4 added to the I column. Trump gains 3.5 from the 7.4 added to the I column because Biden won Indys in this survey, but only barely. Adjusted result, Biden 43.5, Trump 45.5.
1048RV, (Their likely voter sample is not real, just a slightly smaller RV sample). 32D/33I/27R; Leaners 49D/6I/43R.
49
38
x
Political Ref's Take on the above poll
The D+6 sample oversamples Dems and there are too many I's, will adjust to 34D/30I/33R. This is also a registered voter poll so a registered voter screen will apply to the I column. Biden had a 6-point partisan advantage and an 11-point lead, suggesting he did about 5 (15% across three parties) points better across party, assuming Biden won D's 93-3, I's 50-38 (43-35) and Trump won R's 88-9. Applying the new ratios to the more likely turnout, Biden receives 47.5 to Trump 40.5.
1048RV, (Their likely voter sample is not real, just a slightly smaller RV sample). 32D/33I/27R; Leaners 49D/6I/43R.
49
38
x
Political Ref's Take on the above poll
The D+6 sample oversamples Dems and there are too many I's, will adjust to 34D/30I/33R. This is also a registered voter poll so a registered voter screen will apply to the I column. Biden had a 6-point partisan advantage and an 11-point lead, suggesting he did about 5 (15% across three parties) points better across party, assuming Biden won D's 93-3, I's 50-38 (43-35) and Trump won R's 88-9. Applying the new ratios to the more likely turnout, Biden receives 47.5 to Trump 40.5.
1048RV, (Their likely voter sample is not real, just a slightly smaller RV sample). 32D/33I/27R; Leaners 49D/6I/43R.
49
38
x
Political Ref's Take on the above poll
The D+6 sample oversamples Dems and there are too many I's, will adjust to 34D/30I/33R. This is also a registered voter poll so a registered voter screen will apply to the I column. Biden had a 6-point partisan advantage and an 11-point lead, suggesting he did about 5 (15% across three parties) points better across party, assuming Biden won D's 93-3, I's 50-38 (43-35) and Trump won R's 88-9. Applying the new ratios to the more likely turnout, Biden receives 47.5 to Trump 40.5.
1048RV, (Their likely voter sample is not real, just a slightly smaller RV sample). 32D/33I/27R; Leaners 49D/6I/43R.
49
38
x
Political Ref's Take on the above poll
The D+6 sample oversamples Dems and there are too many I's, will adjust to 34D/30I/33R. This is also a registered voter poll so a registered voter screen will apply to the I column. Biden had a 6-point partisan advantage and an 11-point lead, suggesting he did about 5 (15% across three parties) points better across party, assuming Biden won D's 93-3, I's 50-38 (43-35) and Trump won R's 88-9. Applying the new ratios to the more likely turnout, Biden receives 47.5 to Trump 40.5.
1048RV, (Their likely voter sample is not real, just a slightly smaller RV sample). 32D/33I/27R; Leaners 49D/6I/43R.
49
38
x
Political Ref's Take on the above poll
The D+6 sample oversamples Dems and there are too many I's, will adjust to 34D/30I/33R. This is also a registered voter poll so a registered voter screen will apply to the I column. Biden had a 6-point partisan advantage and an 11-point lead, suggesting he did about 5 (15% across three parties) points better across party, assuming Biden won D's 93-3, I's 50-38 (43-35) and Trump won R's 88-9. Applying the new ratios to the more likely turnout, Biden receives 47.5 to Trump 40.5.
Political Ref's Poll Average
