Here are my updated crude 31- and 14-day polling averages. I exclude GOP troll polls (Rasmussen, Trafalgar, etc), any polls released by partisan organizations, and some with really sketchy methods such as ActiVote and Gateway. For the 14-day state averages, the number of polls is in parentheses.
Here are my updated crude 31- and 14-day polling averages. I exclude GOP troll polls (Rasmussen, Trafalgar, etc), any polls released by partisan organizations, and some with really sketchy methods such as ActiVote and Gateway. For the 14-day state averages, the number of polls is in parentheses.
AZ: 31-day T+1.4, 14-day T+1.5 (6)
GA: 31-day T+1.0, 14-day T+1.4 (7)
MI: 31-day H+1.0, 14-day H+0.9 (11)
NV: 31-day H+1.2, 14-day T+0.3 (4)
NC: 31-day T+0.5, 14-day T+0.3 (6)
PA: 31-day H+0.8, 14-day H+0.4 (11)
WI: 31-day H+1.0, 14-day H+0.1 (8)
US: 31-day H+3.0, 14-day H+2.6
These numbers aren't as good for Harris as they were earlier this month. The frequency of state polling has dropped except in MI and PA and I would take the 31-day averages more seriously than the 14-day ones. For example, WI's 14-day average no longer includes the Marquette poll showing Harris up 4. The 14-day average is more meaningful for the US. A number of national polls have shown movement toward Trump lately, but as Dies Irae noted yesterday this seems to be mainly coming from shifts in the party ID of their samples, in many cases showing R+3 national samples which would match the 2022 midterm and would seem to be too red for a general election in which Dem enthusiasm seems to be at least as high as it was in 2020. I suspect the recent red shift reflects change in response bias more than change in voting intentions.
There's a gap in high-quality state polling recently, likely due to the calendar; pollsters want to get their final reads in late October and don't usually do polls in the week or two before then. Unless they're crappy pollsters (cough cough red wavers).
Here are my updated crude 31- and 14-day polling averages. I exclude GOP troll polls (Rasmussen, Trafalgar, etc), any polls released by partisan organizations, and some with really sketchy methods such as ActiVote and Gateway. For the 14-day state averages, the number of polls is in parentheses.
AZ: 31-day T+1.4, 14-day T+1.5 (6)
GA: 31-day T+1.0, 14-day T+1.4 (7)
MI: 31-day H+1.0, 14-day H+0.9 (11)
NV: 31-day H+1.2, 14-day T+0.3 (4)
NC: 31-day T+0.5, 14-day T+0.3 (6)
PA: 31-day H+0.8, 14-day H+0.4 (11)
WI: 31-day H+1.0, 14-day H+0.1 (8)
US: 31-day H+3.0, 14-day H+2.6
These numbers aren't as good for Harris as they were earlier this month. The frequency of state polling has dropped except in MI and PA and I would take the 31-day averages more seriously than the 14-day ones. For example, WI's 14-day average no longer includes the Marquette poll showing Harris up 4. The 14-day average is more meaningful for the US. A number of national polls have shown movement toward Trump lately, but as Dies Irae noted yesterday this seems to be mainly coming from shifts in the party ID of their samples, in many cases showing R+3 national samples which would match the 2022 midterm and would seem to be too red for a general election in which Dem enthusiasm seems to be at least as high as it was in 2020. I suspect the recent red shift reflects change in response bias more than change in voting intentions.
There's a gap in high-quality state polling recently, likely due to the calendar; pollsters want to get their final reads in late October and don't usually do polls in the week or two before then. Unless they're crappy pollsters (cough cough red wavers).