At 01:19 PM 2/19/2004, you wrote:
>18% of 10,000 is 1,800 and that should produce significant results for
>even minute differences.
I was talking about opinion polls which typically work with 1000=N (or less). 18% of 1000 is 180.
> One of the problems working with large samples
>is that almost every difference is significant.
>A common way of handling relatively small sub-populations is
>oversampling (and then weighting) - which allows you to further
>disaggregate those population (GSS does that routinely).
duh. but Yoshie was talking about typical 1000=N surveys (they tend not to oversample) and _I_ was talking about 1000=N surveys _I_ was using. (I was playing off another thread, the one that started thsi race baiting/accusations)
>The main point I argued, however, had little to do with the description
>of populations but with some views I encountered without claiming that
>they represent those of a larger population.
i hadn't responded to you, but to doug. that help? and it was a joke, an elbow nudging doug, and i also pushed Yoshie's foot at the same time: she was trying to use a statistical concept to justify a poor argument (stat. significance) without really understanding them.
BTW, Grant, very excellent post on the related thread.
oh and i've gotta go butch up with some rusty iron, so i'll get back to diane and yes! joanna shortly!
kelley