MAINE   POLLS: Gideon +2.3, Gideon 47.0, Collins 44.7 REF'S AVG: Collins +3.4, Gideon 45.7, Collins 49.1 Explanation of Decision: In the crosstabs, Collins is doing well with Democrats and doing slightly better than breaking even with Independents after I apply an RV filter to the I column where necessary. Before that she's doing about even or very slightly behind with I's. I think the party turnout will be even. With Collins doing slightly better across party and at least even with I's, she should win with a big enough margin to offset the ranked choice voting. I also think there is a small hidden vote for Trump in Maine, smaller than other states, but it's there. That will help Collins a bit as well. RETURN TO SENATE RACES 2016 Exit Polls - 31D/39I/30R without leaners (PDF) 2018 Exit Polls - 41D/18I/41R with leaners (PDF) Pollsters' Averages Chart Political Ref's Averages Chart Poll Date Sample + Party Distribution Result Pollster Avg Ref's Avg Emerson College 10/29-10/31 611LV, 36D/37I/27R Gideon 48, Collins 46 x   Political Ref's Take on the above poll D+9 is too high. Assuming 33D/38I/33R. Inferred ratios, Giden leads by 2 w/ 9-point partisan advantage, suggesting Collins gains 7-points across party. Based on other polling, Collins is likely doing better w/ D's than Gideon with R's. Assuming Gideon 80-13D's, 41-45I's, Collins 85-10R's. Applying ratios to more likely turnout, Gideon 45.3 to Collins 49.4. Gideon 45.3, Collins 49.4   x Survey USA 10/23-10/27 1007LV, Not provided Gideon 46, Collins 45 x   Political Ref's Take on the above poll Data: Gideon 81-13Ds, 41-40 I's, Collins 88-8Rs. They assumed 1007 of 1071 RVs contacted are LVs. Too high. An RV screen will apply to I column, making it 41-44.3 for Collins. Assuming 33D/38I/33R. Applying ratios, Gideon receives 45.0 to Collins 50.1. Gideon 45.0, Collins 50.1   x Colby College 10/21-10/25 879LV, 31D/42I/27R Gideon 47, Collins 43 x   Political Ref's Take on the above poll The turnout should be even to a slight R advantage because some Democrats will not have voted by election day and will choose not to go out. That will be a small number, but important. I think even turnout is likely with a lower than normal turnout for I's. Assuming 33D/38I/33R. Turnout wins this one. Big problem, they assumed 879 of 894 RVs contacted were likely voters. This self-selected mode of determning likely voters yields far too many voters. RV screen will apply to I column for this reason. Data shows Gideon wins D's 86-9, I's 42-40 (40-42.3) and Collins win R's 87-10. It's a draw! Until we apply the RV screen, changing I column to 40-42.3 for Collins. Adjusted result when ratios applied to more likely turnout, Gideon 46.9 to Collins 47.8. Gideon 46.9, Collins 47.8   x Colby College 9/17-9/23 847LV, 48D/13I/39R Gideon 45, Collins 41 Repeat   Political Ref's Take on the above poll The D+9 sample oversamples Dems. The leaners sample should be closer to a partisan balance. Will adjust to D+4 leaners, or 45D/14I/41R. Voter ratios are not provided but can be inferred. Gideon had a 9-point partisan advantage but lead by only 4, suggesting Collins did about 5 points (15% over 3 parties) better among Democrats and Independents than Gideon did with Republicans and Independents. Assuming Gideon won D's 90-6, Collins won I's 48-38 and R's 92-3. Applying the ratios to the more likely turnout, Gideon receives 47.1 to Collins 47.1. Gideon 47.1, Collins 47.1   Repeat Suffolk Univ 9/17-9/20 500LV, 38D/32I/30R Gideon 46, Collins 41 >1month   Political Ref's Take on the above poll The D+8 party distribution oversamples Democrats. Will adjust to D+1, matching 2016, or 31D/39I/30R. According to the poll data, Gideon won D's 88-7, Collins won I's 42-35 and R's 86-4. Applying the voter ratios to the more likely turnout, Gideon receives 42.1 to Collins 44.4. Gideon 42.1, Collins 44.4   >1month New York Times/Sienna College 9/11-9/16 663LV, 37D/34I/29R Gideon 49, Collins 44 >1month   Political Ref's Take on the above poll The D+8 sample oversamples Democrats. Will adjust to D+1 from 2016, or 31D/39I/30R. According to the poll data, Gideon won D's 92-5, I's 50-42 and Collins won R's 88-8. Applying the voter ratios to the more likely voter turnout, Gideon receives 50.4 to Collins 44.3. Gideon 50.4, Collins 44.3   >1month Quinnipiac Univ 9/10-9/14 1183LV, 28D/41I/25R Gideon 54, Collins 42 >1month   Political Ref's Take on the above poll The D+3 sample slightly oversamples Democrats. The bigger problem with this poll is the failure to identify how many people were called. It is labeled a likely voter poll but it is of "self-identified" likely voters, which typically means 97-99% of registered voters are included, rendering it effectively a registered poll. A registered voter flter will be applied to the I column. According to the poll data, Gideon won D's 93-6, I's 59-36 (42-23.3) and Collins won R's 89-7. Included   >1month Citizen Data 9/4-9/7 600LV, 35D/37I/31R Gideon 49, Collins 41 >1month   Political Ref's Take on the above poll The partisan split should break down evenly in Maine. Will adjust to 33D/37/33R. Voter ratios are not provided but can be inferred. Gideon leads by 8 with a +4 partisan advantage, suggesting she performed about 4 points (12%) better with Independents. Assuming each candidate won her party 90-5 and Gideon won Independents 50-38, Gideon receives 49.9 to Collins 45.4. Gideon 49.9, Collins 45.4   >1month Fabrizio Ward/Hart Research 8/30-9/5 800LV, 40D/28I/32R Gideon 44, Collins 43 >1month   Political Ref's Take on the above poll The party distribution is a bit high on the D side. Will adjust to an even partisan split of 36D/28I/36R. According to the poll data, Gideon won D's 84-13, Collins won I's 52-41 and R's 86-10. Applying the voter ratios to the more likely voter turnout, Gideon receives 45.3 to Collins 50.2. Gideon 45.3, Collins 50.2   >1month PPP 8/13-8/14 571RV, 38D/32I/30R Gideon 49, Collins 44 >1month   Political Ref's Take on the above poll The D+8 sample oversamples Democrats. Will adjust to the much more likely D+1, or 31D/39I/30R. Voter ratios are not provided but can be inferred. Gideon had a 8 point party turnout advantage but leads by 5, so Collins did better with Independents. Will assume Collins won Independents 16.5-13.5, or 52-42%. Assuming each candidate wins here party 90-5, the final adjusted result is Gideon 45.8 to Collins 48.8. Gideon 45.8, Collins 48.8   >1month RMG Research 7/27-8/2 500RV, Party Distribution Not Provided Gideon 48, Collins 41 >1month   Political Ref's Take on the above poll This is a registered voter poll and a registered voter screen will be applied. On average 31% of registered voters will not vote, 55% of which are Dem leaning and 41% are GOP leaning. Adjusted, Greenfield receives 31 to Collins 28.3. Applying the 31/28.3 ratio to the 89% of respondents who chose one of the two women, Greenfield receives 46.5 to Collins 42.5. Gideon 46.5, Collins 42.5   >1month Quinnipiac Univ 7/30-8/3 807RV, 29D/39I/25R Gideon 47, Collins 43 >1month   Political Ref's Take on the above poll The D+4 sample oversamples Democrats. Adjusting to the more likely D+1 sample, or 31D/39I/30R. Will also apply a registered voter screen to the Independent column. According to the poll data, Gideon won D's 88-7, I's 46-42 and Collins won R's 85-8. Adjusting I's to 29 for Greenfield and 29.3 for Collins. Applying these ratios to the more likely voter turnout, Gideon receives 41.0 to Collins 39.1. Gideon 41.0, Collins 39.1   >1month Data for Progress 7/24-8/2 866LV, Party Distribution Not Provided Gideon 48, Collins 45 >1month   Political Ref's Take on the above poll Because this is a leaners poll result, I will use the leaners voter turnout from 2018 of 41D/18I/41R. According to the poll data, Gideon won D's 89-10, I's 49-35 and Collins won R's 88-8. Applying these ratios to the likely turnout, Gideon receives 48.6 to Collins 46.5. Gideon 48.6, Collins 46.5   >1month Colby College 7/18-7/24 888LV, 31D/41I/27R Gideon 44, Collins 39 >1month   Political Ref's Take on the above poll Will adjust to the more likely voter turnout of 31D/39I/30R. According to the poll data, Gideon won D's 79-12, I's split 38 for Gideon to 35 for Collins, and Collins won R's 75-13. Applying these ratios to the more likely turnout, Gideon receives 43.2 to Collins 39.9. Gideon 43.2, Collins 39.9   >1month PPP 7/22-7/23 561RV, 38D/33I/29R Gideon 47, Collins 42 >1month   Political Ref's Take on the above poll The D+8 sample oversamples Democrats. Adjusting to the 31D/39I/30R 2016 voter turnout model. Voter ratios are not provided but can be inferred. Collins had a -9 voter turnout but trails by only 5 points, so she must have done better with Democrats and Independents than Gideon did with Republicans and Independents. Assuming Collins won I's by 4, or 18.25-14.25 or 56% to 44%. I will assume each candidate won their party 90-5, although this is likely not the case it works for purposes of calculation. Applying the voter ratios to the more likely turnout, Gideon receives 46.6 to Collins 50.4. Gideon 46.7, Collins 50.4   >1month PPP 7/2-7/3 1022RV, 38D/32I/30R Gideon 46, Collins 42 Repeat   Political Ref's Take on the above poll The D+8 sample oversamples Democrats. Adjusting to the 31D/39I/30R 2016 voter turnout model. Voter ratios are not provided can be inferred. Collins had a -8 voter turnout but trails by only 4 points, so she much have done better with Democrats and Independents than Gideon did with Republicans and Independents. Assuming Collins won I's by 4, or 18-14 or 56% to 44%. I will assume each candidate won their party 90-5, although this is likely not the case it works for purposes of calculation. Applying the voter ratios to the more likely turnout, Gideon receives 46.6 to Collins 50.4. Gideon 46.7, Collins 50.4   Repeat Moore Information as reported by fivethirtyeight.com 6/20-6/24 600RV, Party Distribution Not Available Gideon 37, Collins 45, Savage 3 as reported by fivethirtyeight.com >1month   Political Ref's Take on the above poll Neither party distribution or voter ratios within parties was provided. Adjustment not possible. Adjustment not possible.