Question wording affects answers–and it must be kept in mind that it’s impossible to word questions so that questions do not affect/bias/skew answers. For example, the Chicago Tribune (11/21/15, sec. 1, p. 3) reports that “the U.S. Census Bureau is considering changes . . . that would reclassify some minorities who were considered `white’ in the past,” thereby decreasing the percentage of people who say they are white and shorting the time when whites become the minority.
“In the past, `white’ was the only . . . option available to Arab-American respondents. . . . [and] for many years, many U.S. Latinos also checked the `white’ box because options were limited”. However, to what extent, if any, Arab-Americans and Latinos will opt out of the “white” box and check “Arab-American” and “Latino” is unknown; both–again to what extent, if any–may be reluctant to say who they are because they “fear the federal government”.
When you want to find out what’s really going on, don’t ask. That’s the theme of my book, The Problem with Survey Research, wherein I also briefly discuss observation, experiments, and other procedures that, in contrast to survey research, produce reliable information.