Summary: The degree to which respondents understand survey items and the extent to which they respond truthfully are highly significant issues in assessing the validity of survey results. Focus group interviewing and randomized response procedures are two techniques used in social science research which tend to minimize these problems. Focus group interviewing is a means for developing the information needed to match the respondent understanding level to a survey instrument and to increase the information known about the survey topic prior to questionnaire design. A focus group interview can result in the kinds of information needed to develop a workable data collection instrument. It is a quasi-structured interview with a group of respondents chosen to typify a variety of anticipated survey respondents. Focus group analysis information should be added to the knowledge that the researcher already has. In many cases, it will reinforce questionnaire design decisions previously made. In other cases, it may suggest different question wordings or entirely different lines of questions than had been originally envisioned. Randomized response is used to increase respondents' willingness to answer sensitive questions truthfully by providing the respondent complete confidentiality. A sensitive question is paired with a nonsensitive question with a known aggregate response pattern. Among the drawbacks of randomized response are that: respondent characteristics cannot be linked with an individual's answers to sensitive questions; the relationship between the incidence of admitted behavior and actual behavior is usually unknown; and it does not work well with less sophisticated respondents. Randomized response seems most useful for assessing the incidence of antisocial behaviors or feelings.
