n the experiment, two separate random samples of more than 1,000 registered voters were asked to read a profile sent to them online of a hypothetical candidate for u.s. congress in their district. one random sample of 1,161 respondents read a profile of ann clark, described as a lawyer, a churchgoer, a member of the local chamber of commerce, an environmentalist and a member of the same party as the survey respondent. they were then asked what they liked and didn't like about her, whether they considered her qualified and whether they were inclined to vote for her. there was no indication that this was a survey about gender or gender bias. a second random sample of 1,139 registered voters was asked to read a profile of andrew clark, who—except for his gender—was identical in every way to ann clark. these respondents were then asked the same questions.