Recent research reveals that social-desirability bias remains active in the measurement of white anxieties about the changing racial composition of the country:
First, we asked respondents to tell telephone interviewers whether they agreed or disagreed with the statement, “The idea of an America where most people are not white bothers me.” Among whites, 13 percent admitted to an interviewer that the idea of a majority-minority America bothers them. There was only modest variation among white subgroups, ranging from 10 percent of younger whites young than 50 years of age at the low end to 18 percent of white Republicans at the high end who said an America that is not mostly white concerns them.
Next, we employed a technique called a list experiment, which is designed to allow respondents to indirectly express their views on sensitive subjects. We divided the survey respondents into two demographically identical groups and asked each group to tell us how many, but not which specific items from a list bothered them. One group was designated as a control group and received three control statements, while the other group was designated as a treatment group and received the same three control statements plus a fourth statement that read, “An America that is not mostly white.” Because the control and treatment groups were demographically identical, any variation in the average number of statements chosen between the groups is solely attributable to respondents in the treatment group picking the treatment statement. For any subgroup (but not for an individual), then, one can statistically estimate the proportion of respondents choosing the treatment statement by subtracting the mean number of statements chosen by the treatment group from the mean number of statements chosen by the control group. That number is presented in the chart below as the “indirect response.”