CSB Working Paper, 2021, 9: Using a factorial survey to estimate the relative importance of well-being dimensions according to older people: insights from a repeated survey experiment in Flanders

In this paper, we investigated the potential of a factorial survey to estimate the relative importance of the well-being dimensions of health, income, social relations, leisure and religion or spirituality, according to the views of older people. For this purpose, a factorial survey was implemented in a longitudinal online survey among 800 older adults in Flanders (the Flemish region of Belgium). The potential of the factorial survey was explored in two ways. First, we performed several within-s…

In this paper, we investigated the potential of a factorial survey to estimate the relative importance of the well-being dimensions of health, income, social relations, leisure and religion or spirituality, according to the views of older people. For this purpose, a factorial survey was implemented in a longitudinal online survey among 800 older adults in Flanders (the Flemish region of Belgium). The potential of the factorial survey was explored in two ways. First, we performed several within-sample test-retests to investigate the consistency of the estimated relative importance weights over time (i.e., temporal reliability). Second, we tested the feasibility of the factorial survey for the target population by studying two indicators of cognitive load: response time and response consistency. Overall, we found evidence that the factorial survey works well among a sample of older people. Health, income and social relations were the most important well-being dimensions – followed by leisure and engaging activities. Religion or spirituality appeared to be rather unimportant. The results proved to be consistent in most of the test-retest analyses. In addition, we found that respondents were able to cope well with the complexity of the factorial survey and produced a high level of response consistency within an acceptable amount of response time.