LIKELY voters doesn’t change your percentages.
If someone is determined to be an unlikely voter they are dropped from the poll.
Likely voters do not change the percentages, especially 10%..
This is the breakdown of WHO voted by part in 2020 in PA:
D 4002630
R 3462524
I 1300016
TOTAL 8765170
D -45.6
R - 39.4
I - 14.8
or as I have posted again and again... Almost exactly in line with registration percentages... D -46, R 39 and I 15
Tracking down likely voters is determining if the voter is LIKELY to vote, it doesn’t change the percentages that vote in relationship to registration numbers.. I don’t know who started that lie, but that’s an abject LIE.
Voter turnout by party mimics their registration percentages insanely closely, especially in large statewide races in all but the rarest of exceptions.
Polling likely voters, does NOT change the turnout of percentages of each, it just means they poll made efforts to weed out folks they determined were unlikely to vote. The goal should still, and always be that the polling sample by party be in line with the registration.
Trafalgars polls this cycle in PA have been grossly oversampling republicans and undersampling Independents... they have also consistently been oversampling women as well.
Numbers come directly from the PA Department of State:
” if the voter is LIKELY to vote, it doesn’t change the percentages that vote in relationship to registration numbers”
OF COURSE IT DOES! If Republicans are more likely to come out to vote than Democrats, you need to sample a higher percentage of them compared to registration. That is what defines their greater enthusiasm.
Suppose you have 50:50 split between Republican and Democrat, but Republicans are more enthusiastic. The people who then CAST VOTES will be maybe 53:47 Republican to Democrat.