Question order sensitivity of subjective well-being measures: focus on life satisfaction, self-rated health, and subjective life expectancy in survey instruments

This study examines the effect of question context created by order in questionnaires on three subjective well-being measures: life satisfaction, self-rated health, and subjective life expectancy. We conducted two Web survey experiments. The first experiment (n = 648) altered the order of life satisfaction and self-rated health: (1) life satisfaction asked immediately after self-rated health; (2) self-rated health immediately after life satisfaction; and (3) two items placed apart. We examined their correlation coefficient by experimental condition and further examined its interaction with objective health. The second experiment (n = 479) asked life expectancy before and after parental mortality questions. Responses to life expectancy were compared by order using ANOVA, and we examined interaction with parental mortality status using ANCOVA. Additionally, response time and probes were examined. Correlation coefficients between self-rated health and life satisfaction differed significantly by order: 0.313 (life satisfaction first), 0.508 (apart), and 0.643 (self-rated health first). Differences were larger among respondents with chronic conditions. Response times were the shortest when self-rated health was asked first. When life expectancy asked after parental mortality questions, respondents reported considering parents more for answering life expectancy; and respondents with deceased parents reported significantly lower expectancy, but not those whose parents were alive. Question context effects exist. Findings suggest placing life satisfaction and self-rated health apart to avoid artificial attenuation or inflation in their association. Asking about parental mortality prior to life expectancy appears advantageous as this leads respondents to consider parental longevity more, an important factor for true longevity.