...differ?

There are three response patterns we analyze: whether a respondent knew or not, accuracy among knowers, and accuracy among guessers.

...positions.

In a previous piece (Alvarez and Gronke 1996), we examined citizen awareness of their representatives' votes (both House and Senate) on the Persian Gulf War resolution. We found that citizen awareness of how their representatives voted, while not great, was not trivial. Citizen ability to ``recall" or ``guess" their representatives' positions is structured by individual characteristics and a reasonable set of contextual cues.

...Study.

All of the results we present in this paper are from a pooled sample of Senators. We simply created a dataset where there can be two observations per respondent, with one observation being their responses for one senator and the other observation being their responses for the other senator. This is the same analysis strategy we followed in our earlier work (Alvarez and Gronke 1996).

...coefficient.

The marginal effects are estimated as the difference in probability achieved by by moving a continuous variable one unit from the sample mean. Alternatively, for binary variables, the marginal effects are estimated as the difference in probability produced by moving the binary indicator from zero to one.

Paul Gronke
Sun Nov 24 22:06:23 EST 1996