The evidence - evaluative measures
Rather than specifying a particular reference period, evaluative questions typically ask respondents how their life is overall nowadays or these days (examples at Annex A). Thus, whilst inviting an assessment of life that might span the entire life-course, respondents are asked to evaluate their life at this point in time, which is likely to favour the recent past (Diener, Inglehart and Tay, 2012). Cognitive testing work undertaken by the UK’s ONS in the course of developing experimental questions on subjective well-being (ONS, 2011c) indicated that “nowadays” was interpreted by respondents in several different ways, ranging from the present moment, the last five years, or since key life events. “Nowadays” was also considered to be a dated or old-fashioned term. The term “these days” was meanwhile considered as referring to a shorter time-frame (e.g. the past few weeks).
It is difficult to identify definitively a right or wrong reference period for evaluative measures. As Diener, Inglehart and Tay (2012) note, there has thus far been little research into i) how the framing of life satisfaction questions influence scores; and ii) which framing is most valuable for policy. For data to be comparable, consistency in the use (or not) of reference periods is certainly preferable - although it is not at present clear how important this consistency is, and how precisely reference periods need to match. Some of the most widely-used evaluative measures in the literature at the moment ask at present (Gallup World Poll) and these days (World Values Survey) - but it seems unlikely that these two terms should produce large differences in responses. Greater variability may be found between those questions that use a reference period and those that do not, although again the evidence regarding the precise size of the impact this may have is lacking.