In market research, a common need is to assess preference or importance for a list of items, usually characteristics of a product or service.
Typical approaches are rating or ranking scales, but they present important drawbacks. Rating scales results are typically associated with a lack of differentiation (i.e., all items are rated as similarly important) and with a different usage of the scale among different respondent segments (e.g., due to cultural or experience bias). While ranking scales are not affected by these issues, they are collected on an ordinal scale (as opposite as to a interval or ratio scale) which does not allow robust statistical analyses to be performed. Also, it is typically challenging and time consuming to assess a large number of items through a ranking approach (full or partial rank).
A valid solution to assess the value customers attach to a list of features consists in adopting a trade-off approach. By far, the two most popular ones are:
- Paired Comparisons of Statements (PCS) and
- Maximum Difference Scaling (MDS or maxdiff).
PCS is to some extent similar to MDS, with the main difference being its exercise, simpler and thus more appropriate for some types of respondents (the elderly) or for some interview modes (e.g., CATI).