The unintended effects of survey design and interviewer performance threaten the validity of every survey. To minimize the threat to this survey, the research team drew on a great deal of expertise in questionnaire design and put extraordinary effort into interviewer training. Unfortunately, such measures never entirely succeed. We have therefore tested for several potential artifacts that might give rise to unwarranted conclusions.

Four of our tests bear on special features of the Soviet Interview Project (SIP). Briefly, the first one examines the extent to which responses were affected by the time lapse between respondents' departure from the Soviet Union and the date of their interview. Since the survey focused on respondents' last normal period in the Soviet Union, we would like to identify any variables affected by deteriorating recall or contamination by Western experience.

The second test searches for differences between answers from willing and reluctant respondents. Of course, including data from reluctant respondents did not in itself render this survey unique, for all respectable survey organizations endeavor to convert “soft refusals” in the effort to achieve high response rates. Nor were most of the reasons offered by reluctant respondents unusual: that they did not know enough to be worthy of an interview or that they lacked time or interest. A few members of the sample, however, voiced a unique concern: fear that answers might bring reprisals to relatives in their homeland or unforeseen consequences in the United States.