Correction on “Polling Shenanigans?”
Oops. A reader familiar with polling methods and jargon corrects my misunderstanding of the number of black respondents in the poll:
What the Post did with the extra interviews of blacks is called “oversampling,” and … there’s nothing wrong with it. If the Post is doing what pollsters normally do, all they’ve done is interview extra African-Americans for the purpose of lowering the margin of error for only the African-American portion of the sample. In other words, the results WaPo is reporting for ”adults” only consists of results from the sample 1,125 adults, and that sample contains a percentage of African-Americans proportionate to the population as a whole; the purpose of the extra interviews of African-Americans is to enable the pollsters to have a sample of African-Americans that is large enough so that they can have a meaningfully report results just for that population segment.…
This same reader adds this insight:
I think the real point to be made about the WaPo poll results is that, notwithstanding the addition of Roberts and Alito to the Supreme Court, it appears an overwhelming majority of Americans believe that the Court is either just right or too liberal.… And what’s so surprising about the fact that 31% think the Supreme Court is too conservative? Hillary Clinton or Barack Obama, for example, would get at least that percentage in a general election.