This makes sense to me. Let's hear from the people who are still persuadable first, not from the two groups who have made up their minds. What news is affecting the persuadables?
Re Judith and Matthew, even though Dick pushes back somewhat against this, and as I can say as a "non-duopoly" voter, most those "independents" are actually pretty much R's or D's who don't want party labels, as I see it. (Asking the right questions, like drilling down on Israel-Gaza enough, or the War on Drugs, you'll find that out, too.)
Speaking of, more polls should include possible third-party/real independent voters and questions about 3rd/I candidates in a year like this. (Not doing so is part of how polls become self-fulfilling prophecies.) They should ask WHY they're not voting for one of the duopoly candidates.
This is a good idea in concept, but polls are designed and weighted to be representative of the topline, not the subgroup. That's why we see subgroups swing around, including independents - remember October 2022 movement in the NYT poll toward Republicans? That drove a lot of red wave fears, and turned out to be completely false.
if polls are designed and weighted to represent independents, then absolutely yes. But as it is, this should not be the lede because its subject to all the subgroup problems.
Also, it is very important how many Democrats Biden is losing, for example. If he holds 95% of Dems and Trump holds 88% of Republicans, the independents are not telling the whole story. In polls now, Biden is losing considerable proportions of Dems, which is a huge concern. (and see also caveats about subgroups, because Ds & Rs are also subgroups)
Good insights, as usual, Dick. Allow me to add a thought or two. Lots has been written about the fallibility of poll results, starting with the theory that many pollsters base their sampling on the prior election not the prospective one (thus baking in preconceptions) and whatever corrective adjustments they try to make are based on educated guesses as likely to be wrong as not. In that regard political reporters are probably well advised to take their own measure of the electorate and, more to the point, believe their eyes. Years ago while his colleagues clambered on and off planes and buses to hear a candidate’s stump speech for the umpteenth time, David Broder used to knock on doors and chat with voters. Scientific? Hardly, but added to instinct tempered by decades of experience it gave him a feel for what was really afoot. I thought of that in 2016 when it was only AFTER the election that reporters in Michigan and elsewhere began to remember the inordinately high number of Trump yard signs they’d seen and the relative speed with which the Trump campaign had produced enormous crowds.
The polls and resultant poll stories have two major problems in my view.
First, self-identifying as an independent does not autmatically mean that you are politically neutral. Anecdotal, but many "independents" I know almost always vote left or almost always vote right. They like to maintain the illusion of their own impartiality. I've never seen good numbers on how many "independents" are left leaning or right leaning a priori their choice of candidate. I suspect the self ID'd independents skew right but I don't know and neither seemingly does anyone reporting on this.
The second major problem, frequently identified is that national polls in a presidential race simply don't matter. While the polls today show a roughly even national race, if we are to believe the RCP state-by-state averages, the race is currently comfortably in Trump's hands. He wins the electoral college in a near walk. "Poll stories" seldom parse things this closely.
From there, every other number is skewed putting aside the difficult issues of intent to vote, difficulties in reaching cell phones and all of the other issues we all know about.
If we're going to have to have "poll stories" journalism needs to do a much better jo of providing context for the numbers they report.
This makes sense to me. Let's hear from the people who are still persuadable first, not from the two groups who have made up their minds. What news is affecting the persuadables?
I agree with Matthew.
Did not know there were so many independents.
Excellent column.
Most thoughtful and should be must reading for the political and economic press.
Re Judith and Matthew, even though Dick pushes back somewhat against this, and as I can say as a "non-duopoly" voter, most those "independents" are actually pretty much R's or D's who don't want party labels, as I see it. (Asking the right questions, like drilling down on Israel-Gaza enough, or the War on Drugs, you'll find that out, too.)
Speaking of, more polls should include possible third-party/real independent voters and questions about 3rd/I candidates in a year like this. (Not doing so is part of how polls become self-fulfilling prophecies.) They should ask WHY they're not voting for one of the duopoly candidates.
This is a good idea in concept, but polls are designed and weighted to be representative of the topline, not the subgroup. That's why we see subgroups swing around, including independents - remember October 2022 movement in the NYT poll toward Republicans? That drove a lot of red wave fears, and turned out to be completely false.
if polls are designed and weighted to represent independents, then absolutely yes. But as it is, this should not be the lede because its subject to all the subgroup problems.
Also, it is very important how many Democrats Biden is losing, for example. If he holds 95% of Dems and Trump holds 88% of Republicans, the independents are not telling the whole story. In polls now, Biden is losing considerable proportions of Dems, which is a huge concern. (and see also caveats about subgroups, because Ds & Rs are also subgroups)
Good insights, as usual, Dick. Allow me to add a thought or two. Lots has been written about the fallibility of poll results, starting with the theory that many pollsters base their sampling on the prior election not the prospective one (thus baking in preconceptions) and whatever corrective adjustments they try to make are based on educated guesses as likely to be wrong as not. In that regard political reporters are probably well advised to take their own measure of the electorate and, more to the point, believe their eyes. Years ago while his colleagues clambered on and off planes and buses to hear a candidate’s stump speech for the umpteenth time, David Broder used to knock on doors and chat with voters. Scientific? Hardly, but added to instinct tempered by decades of experience it gave him a feel for what was really afoot. I thought of that in 2016 when it was only AFTER the election that reporters in Michigan and elsewhere began to remember the inordinately high number of Trump yard signs they’d seen and the relative speed with which the Trump campaign had produced enormous crowds.
The polls and resultant poll stories have two major problems in my view.
First, self-identifying as an independent does not autmatically mean that you are politically neutral. Anecdotal, but many "independents" I know almost always vote left or almost always vote right. They like to maintain the illusion of their own impartiality. I've never seen good numbers on how many "independents" are left leaning or right leaning a priori their choice of candidate. I suspect the self ID'd independents skew right but I don't know and neither seemingly does anyone reporting on this.
The second major problem, frequently identified is that national polls in a presidential race simply don't matter. While the polls today show a roughly even national race, if we are to believe the RCP state-by-state averages, the race is currently comfortably in Trump's hands. He wins the electoral college in a near walk. "Poll stories" seldom parse things this closely.
From there, every other number is skewed putting aside the difficult issues of intent to vote, difficulties in reaching cell phones and all of the other issues we all know about.
If we're going to have to have "poll stories" journalism needs to do a much better jo of providing context for the numbers they report.