Skip to main content

View Diary: Race and the Tea Baggers: the data we've been waiting for (234 comments)

Comment Preferences

  •  CAUTION 'til survey composition clarified... (0+ / 0-)

    This UW survey was discussed thoroughly in another diary by Hesiod yesterday...

    http://www.dailykos.com/...

    The DKos User, Marigold, added a previously-unknown piece of info on the survey's methodology and sample composition and the following interesting link...

    http://uwnews.org/...

    That link describes the survey's method and sample composition as follows:

    Conducted by telephone from Feb. 8 to March 15, the survey reached 494 whites, 380 blacks, 77 Latinos and 64 members of other races. The sampling error margin is plus or minus 3.1 percentage points. The Washington Institute for the Study of Ethnicity, Race and Sexuality and the UW Department of Political Science paid for the survey. It was conducted by the UW's Center for Survey Research.

    That sample totaling n=1015, if correctly described at the uwnews.com link, is NOT a representative sample of 1015 people. Instead it appears to be either:

    A) a smaller (and perhaps representative) sample of some lower number of people, with an augment of "blacks" to a more readable level (n=380) than would be natural in a sample of 1,000 (where it would be n=about 120). And with n=only 77 "Latinos" (just 7.7% of n=1000 instead of the usual 14-15% for "hispanic/latino/spanish"), the survey is hardly representative of that sub-group.

    B) or, perhaps, it is a stratified sample of "whites" and "blacks" with little attention to other ethnicities.

    In any case, with the non-representative distribution of ethnicities in that sample (IF the UWnews.org link is to be believed), the claimed MOE of +/- 3.1% cannot correct. This is clearly NOT a representative sample of about 1,000 people - for whom a 3.1 MOE would be possible.

    I would caution Kossacks to withhold comment until someone can sort out the facts behind the sample composition.  In the other diary linked above, Deoliver notes that she knows the survey's author and is writing him for clarification on the study. We should wait until De gets more information from him. This might be a valid study, but poorly described in the UWnews.org link; or it might be a poor study with poor MOE description further compounding non-expert survey design.

    Learn more about second-class U.S. citizenship at http://www.equalitymatters.org/

    by Larry Bailey on Fri Apr 09, 2010 at 05:16:35 AM PDT

    [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site