3 comments

  1. This seems like less of an outright lie and more of a misunderstanding of what polls do in the first place. Asking if the poll is a ‘snapshot’ OR ‘hypothesis’ is a false dichotomy. Nearly all polls/surveys are both snapshots and hypotheses – they are snapshots of the sample that is hypothesized to represent the target population. Any poll that is not purely random and which defines its own target population (ie ‘likely’ voters), will run into the same problem. This isnt really all that disingenuous unless you misrepresent how you obtained your sample. The problem is, when faced with results of a poll like this (or any poll for that matter), few people bother to ask how ‘likely voter’ is defind. Or more importantly, few pay attention to the fact that with a 4 point confidence interval, a 4-point lead is pretty unlikely to be statistically significant difference, so this whole result is pretty meaningless anyways.
    Since there are a variety of reasons why the media would like to portray this race as closer than it probably is, I can understand why this guy tries to go after the methodology itself– but its really more of a scientific literacy/interpettaion issue than bad science. His main critique–that Obama supporters make up a disproportionately amt of those thought to be “unlikely voters”–seems to ignore a substantial amount of evidence that newly registered voters are far less likely to actually vote. Obama’s big voter registration push is likely to yield a lot less actual votes for him than it seems to because those who register and claim to support him now are some of the least likley to show up at the polls.

    Reply

Leave a Reply

Your email address will not be published.

CAPTCHA

*