Response Contamination by Third Parties in a Household Interview Survey

The issue of third parties' presence in face-to-face interview situations is a familiar, yet often unexamined phenomenon. Whilst there is often an implicit or explicit instruction to interviewers to interview respondents alone, there is little conclusive evidence to suggest that the presence of third parties introduces bias by influencing respondents' answers. Data from a pilot study for the British Household Panel Study are examined to address the issue of response contamination. Face-to-face interviews with all household members, which included extensive measures of interviewer's perception of interference and influence on responses, allow us to examine three features of response contamination; first, the extent to which a third party was present, the identity of the third party, and whether such interview situations are likely to be dependent on any particular social characteristics; second, whether the presence of a third party had any significant effects on the distribution of responses, and third, the degree of discrepancy between partner's reports. The analysis focuses on substantive topics considered to be especially susceptible to response effects because of their sensitive or subjective nature.