polling question: black voters?

[dougrhess]

I recall being told that black voters (maybe other non-whites, too?) tend to not say they are likely to vote until much later in the season. Is this true? If so, does this affect "likely voter" polls in any way?

[Fmr. Gov. NickG]

Most polls weight their sample by race, so even if this is true, I don't think it would affect the numbers very much, unless these late-deciders vote substantially differently than early deciders.  And since the black vote is realistically pretty monolithic, I don't think there would be that much difference between these groups.

[dougrhess]

Thanks. What do they weight the data to? Past voting patterns based on race, age and gender? Or to the population of citizens? Registeres citizens?

[Fmr. Gov. NickG]

Thanks. What do they weight the data to? Past voting patterns based on race, age and gender? Or to the population of citizens? Registeres citizens?

Typically, exit polling data from recent statewide elections....so they are trying to get the sample to conform to the population that actually voted last time.

[dougrhess]

Oh, well, recent statewide elections would still be off by a large amount for minorities in some states, maybe most, compared to presidential years. No?

[Fmr. Gov. NickG]

Oh, well, recent statewide elections would still be off by a large amount for minorities in some states, maybe most, compared to presidential years. No?

Usually you would only compare presidential years to other presidential years, and vice-versa.

I'm not sure what most polling firms will do in 2006, since there's no data for 2002.

[CollectiveInterest]

Pollsters have to predict who will show-up to vote. This is one of the limitations of polling.