RATING SCALE CONSTRUCTION: USING VERBAL QUALIFIERS
Rohrmann, B., University of Melbourne, Australia
Questionnaires are the dominant data collection method in psychology, mostly using rating scales as primary response mode. A response scale should fulfil psychometric standards of measurement quality as well as practicality criteria, e.g. comprehensibility for respondents and ease of use. Within category scaling, verbal labelling of rating scales has become the prevalent approach to enhancing usability; words are used as "qualifiers" for all points or end-points only. Verbal labelling provides many advantages, such as ease-of-explanation and familiarity; it also facilitates capturing normative judgments. The main disadvantage is inferior measurement quality; also, cultural factors might confound scales, and cross-study comparability is often difficult. It is thus essential to design verbalized scales carefully if equi-distant and unambiguous instruments are to be achieved - yet only few publications provide pertinent information. The principal idea underlying the current research is: to design rating scales using verbal labels which reflect the cognitions of respondents and for which psycholinguistic and psychometric data are available. Therefore a series of studies was conducted to clarify the measurement features of relevant words or expressions in a variety of contexts. In this contribution, quantitative data for 100 verbal scale point labels are presented, covering five qualifier dimensions: Frequency (e.g. often); intensity (e.g. very); probability (e.g. possibly); quality (e.g. good); response to statements (e.g., disagree). The results enable the systematic construction of scales measuring psychological variables and approximating interval scale level. Finally, criteria for choosing words and shaping scale layouts will be outlined.