September 14, 2024

  • el
  • pt
  • Stanford ReCAL Poll Info

    Interesting: Schwarzenegger ahead of Bustamante in Internet survey on California recall vote. My first reaction is "Internet poll? Can't be reliable." However, this is a Stanford University poll conducted by some serious polisci types, including Morris Fiorina (the polisci geeks in the audience will no doubt recognize the name).

    I am curious as to the exact methodology, but the poll does claim to capture a representative sample. My meager statistical skills lead to wonder how one does this with an internet based poll. Presumably there is some sort of stratified sampling or cohort sampling in the process, but it would seem that there would still be populations that would not be adequately sampled.

    At any rate, here're the numbers:

    Should Gray Davis be recalled as governor?
    Yes 58% (493)
    No 42% (355)

    If the governor is recalled, the candidate who receives the most votes will be elected as the new governor. Please select the candidate that you are most likely to vote for.
    McClintock 6%
    Huffington 2%
    Bustamante 24%
    Schwarzenegger 37%
    Ueberroth 4%
    Other (Write-ins) 14%

    All the numbers are here: Appendix to the Stanford University/Knowledge Networks California Ballot Survey.

    The press release is here: Schwarzenegger ahead of Bustamante in Internet survey on California recall vote.

    Hat Tip to: The Political Times

    Posted by Steven Taylor at September 14, 2024 08:13 AM | TrackBack
    Comments

    Hmm. KN doesn't really do "Internet polling." What they do do, however, is rather interesting (and, for some reason, not explained properly on their web site).

    I probably should write up a full blog post on what Knowledge Networks does and what Harris Interactive (who do "Internet polling") does, including digging out the Political Analysis piece that compared their methods with RDD (I think I know where it is).

    Incidentally, I'm surprised Harris hasn't done a California poll of their own. Maybe they don't have the correct propensity weights for a California sample (or they're suffering from severe panel attrition).

    Posted by: Chris Lawrence at September 14, 2024 05:26 PM
    Post a comment









    Remember personal info?