A Word On Pew's Survey

Alec Tyson, a research assistant at Pew, e-mails:

We enjoyed seeing your analysis of our recent May political survey on your blog; one methodology comment regarding your post on young voters and demographics which may be helpful to you in analyzing our polls going forward: It’s true that we interviewed 88 registered voters, ages 18-29, which – as you note – is less than 8% of our unweighted registered voter sample. However, in all of our surveys, we weight our data to census population estimates.

For example, in the May survey, we reached a disproportionately small number of young voters, but after weighting to census data they accounted for 14% of registered voters -- instead of under 8% -- in our weighted sample. Similarly, we reached 364 voters over the age of 65 (about 29% of our unweighted sample), but after weighting they represent 19% of the registered voters. All our analyses are based on weighted, rather than unweighted data.