Poli psy isn't really a new field. Since 1972 the National Election Studies (NES) have analyzed voter decision making by borrowing tools and theories from psychology instead of traditional political science. But as November draws ever nearer, psychological interpretations of why voters go to the polls, how campaigns influence their actions and what makes great presidents are growing.

Some studies seem to show just how shallow many of us are when it comes to picking a candidate--be it for the presidency or the PTA. Consider, for instance, results that Ohio State University's Kira Sanbonmatsu presented on September 1 at the annual meeting of the American Political Science Association in Washington, D.C.: When all else is equal, about half of voters prefer candidates of a particular gender.

Sanbonmatsu polled 455 randomly selected Ohio residents and found that 63 percent of the women had a gender preference, 62 percent choosing another woman. Fewer men--some 51 percent--had a similar gender bias, although more of them, 68 percent, opted for their own sex. "Voters see gender as one way of predicting a candidate's beliefs and even competency on various issues," she notes.

The study confirmed one of Sanbonmatsu guesses--namely, that liberal voters would probably side with female candidates because women are stereotyped as more liberal. Indeed, democratic participants in the survey were less likely to show a preference for male candidates. "Relying on gender stereotypes may be a reasonable shortcut for voters who lack information about the candidates," Sanbonmatsu adds. "Making logical guesses about who is better able to represent you is not necessarily the same as choosing the male candidate because you are prejudiced against the female candidate."

Were the survey's respondents telling the truth about their prejudices? To test them, Sanbonmatsu set up a hypothetical election, giving participants short biographies of two candidates. Half of the people learned about a candidate named Thomas Brown, the other half, Elizabeth Brown; both Browns had identical qualifications. In the end, voters who preferred female candidates had only a 25 percent probability of choosing Thomas, compared to a 59 percent probability of electing Elizabeth. Says Sanbonmatsu, "Gender is one factor, among others, that affects voters' choice, especially when the voter doesn't know much else about the candidates."

One factor that appears to heavily sway voters' choices--and whether or not they even vote--is not how much they like one candidate, but how much they hate other guy, according to another recent study. Jon Krosnick of Ohio State presented these findings, based on NES surveys conducted between 1972 and 1988, also last week in Washington.

"People who have something to lose by not voting are the ones who will go out and vote. A disliked candidate is seen as a threat," Krosnick says. "This goes against the conventional wisdom in political science, which says the important factor is how much difference there is between how much you like each candidate." Krosnick found that people who like both candidates--even to significantly different degrees--were not as likely to vote as those who actively disliked one or the other.

Krosnick discovered that voters--informed and uninformed alike--tend to form impressions about candidates in much the same way as they do friends, neighbors and co-workers: first impressions count more than information learned later on, voters typically begin with positive impressions of unknown candidates, and negative information has a greater impact than positive feedback.