Key messages on scale length
The optimal number of response categories in rating scales will be informed by a combination of reliability, validity, discriminating power and respondent preferences. Some consideration also needs to be given to enabling comparisons with the existing literature, which includes tried-and-tested response formats. On balance, the evidence broadly seems to favour an 11-point numerical scale for evaluative subjective well-being measures - and this is consistent with some of the current most widely-used approaches (including the Gallup World Poll and the German Socio-Economic Panel). Less is known about the optimal number of response options to offer for eudaimonic and affective measures, and this needs to be considered in combination with other factors that influence response formats, such as survey mode, and whether verbal or numerical scale labels are preferred (discussed below).
It is surprising how little reference the literature makes to the underlying construct of interest when discussing the optimal number of response options to present - even though the initial evidence suggests that this could be important. Life evaluations and eudaimonia measures are often presented to respondents in the form of “bipolar” attitude measures (e.g. completely dissatisfied to completely satisfied or disagree completely to agree completely). The intent behind this is to capture both the direction (negative-positive) and intensity (strong-weak) of feeling. 11-point numerical scales appear to be well-suited to this task, and offer the additional advantage of a scale midpoint, which provides respondents with the option of indicating they fall somewhere in between the two scale end-points, rather than forcing a choice in favour of one side or the other. For single-item scales, including some of the most frequently-used life evaluation measures, 11-point numerical scales also perhaps offer the best balance between scale sensitivity and so much choice that respondents are overwhelmed.
For affect measures, one might be interested in measuring either the intensity of feeling or the frequency with which that feeling occurred. Measures of recently-experienced affect are less like attitude measures, in that one is effectively asking respondents to remember a specific experience or to sum experiences over a specific time period. Affect measures also differ from many evaluative and eudaimonic measures in that they may not be seeking information about the direction of feeling, because it is desirable to measure positive and negative affective states separately (see the section on unipolar versus bipolar measures, below). Finally, affect measures are typically assessed through multi-item measures, which means that scale sensitivity is less strongly determined by any single item. However, there appears to be a lack of research that systematically examines these factors in combination with one another.