The evidence on question comprehension
It is difficult to evaluate the overall evidence on question comprehension, because different measures will perform differently. Question comprehension is usually examined through pre-testing and cognitive testing in the course of scale development, which is rarely reported in detail. Response latencies (i.e. the time respondents take to process the question, then construct and deliver their answers) can indicate how easily respondents can grasp a question and find the information needed to construct a response. Recent evidence (ONS, 2011a) indicates that 0-10 single-item questions on evaluative, eudaimonic and affective subjective well-being can be answered, on average, in around 30 seconds, suggesting that they pose no particular problems to respondents. However, short response latencies could also result if respondents are responding randomly or using mental short-cuts or heuristics to help them answer questions.
Empirical research on the validity of measures perhaps offers the best evidence that respondents have understood question meaning. If survey measures show a strong relationship with real-life behaviours and other non-survey indicators, this suggests that responses are meaningful. As described in Chapter 1, there is considerable evidence to support the overall validity of subjective well-being measures, particularly some of the most frequently-used life evaluation measures. However, this research typically draws on a wide range of different questions, and very few studies have systematically compared the impact of different question wordings under identical conditions.