THE IMPACT OF THE PRESENCE OF OTHERS ON A RESPONDENT'S ANSWERS TO QUESTIONS

Although interviewers in survey research are usually instructedto interview the respondent alone, third parties are a commonenough occurrence to have a potentially serious impact on responses.This article examines the effect of third parties present (spouses,children, and in general) on a respondent's answers. Data fromthe 1994 General Social Survey were used. Results show that,for married respondents, answers on 15 questions about marriage,sexual matters and miscellaneous items are hardly affected bythe presence of a spouse. Differences on gender role items emergeas spurious when proper controls for differences in family andhousehold structure and organization are entered. Among 13 questionsabout child values, sexual matters concerning youths, and otheritems, answers are mostly not influenced by the presence ofa child. When controls are applied, only one difference emerges:Respondents appear less approving of premarital sex when anolder child (6 years or more) is present. Among 13 items onsex, religion, personal evaluations and evaluations of people,only self-reported health is affected by the presence of thirdparties. Overall, the impact of third parties on survey responsesis fairly rare and mostly small. Most apparent impacts of thepresence of others turn out to be spurious. Though they areof limited extent, third-party effects need to be better understood.Specific steps toward this end are suggested.