Jeff JarvisMar 3, 2024
The fucking Times. It follows up yesterday's awful poll with another. That they even asked this question is evidence of the bias--the agenda--in their poll. Who made age an "issue"? The credulous Times falling into the right-wing's projection. This is not journalism. Shameful.
Show threadSouprMattMar 3, 2024

@jeffjarvis So, this is the same poll as yesterday; they are just slowly releasing answers over time and harping on their favorite tune: “Biden is old.” I’d like to point out a big flaw in their poll beyond the chosen questions.

Look at the % of people they polled by population density: 24% city, 41% suburb, 35% rural. The actual population density in America: 31% urban, 55% suburban, 14% rural. They are over-representing rural voters by almost 3 to 1 in this poll. It doesn’t appear they corrected for this at all.

Also, 35% of their respondents are from the “South” region, only 20% from the northeast. Although they don’t specify exactly what that means, this doesn’t appear to be anywhere close to an even distribution.

Finally, 66% of their respondents were white. White people are 57.8% of the population according to the latest census.

In summary, this poll HEAVILY over represents southern rural white voters.

Show threadPaul Wermer, CC BY-NC-SA 4.0Mar 3, 2024
@mls14 @jeffjarvis
Since the President is elected by electoral votes won, not total votes won, I suspect looking at how well the poll captures the electoral vote is more significant than the urban/suburban/rural partition
Show threadSouprMatt
@PaulWermer @jeffjarvis Ok then, 35% of respondents being from the south is even worse. AZ, GA, MI, NV, PA, WS are the biggest swing states. Only one is in the south region - GA, which has a large black population they undercounted. So thanks for pointing out that these results are even more skewed!
Mar 3, 2024 at 5:32pmWeb
Show threadPaul Wermer, CC BY-NC-SA 4.0Mar 3, 2024
@mls14 @jeffjarvis interesting point. If the election is dependent only on swing states, that really changes how polls should be designed if we want meaningful results, rather than the bs horserace calling we see.