Konsekwencje wyboru typu skali odpowiedzi w badaniach ankietowych (Consequences of using different types of rating scales)

The choice of the type of rating scale, made often without careful consideration, has important implications, because respondents are forced to transpose the implicit scale they use to differentiate objects, into a scale provided by the researcher. Analysis of empirical data from five studies showed that: (1) respondents have different preferences as to the length of the scale (some people prefer short {YES, NO} scales, others prefer multipoint scales), (2) the numbers of extreme responses in different measurement techniques correlate highly and positively with each other, so the preference for extreme responses can be regarded as respondent’s response style. The article also presented arguments for four theses: (1) an index constructed from responses on Likert-type scales can be treated as quantitative variables, (2) middle response category should be avoided, however, respondents should be allowed to provide “don’t know” answers, (3) “Don’t know” answer category should appear at the end of the scale - not in the middle, (4) the shape of a distribution of an index constructed from answers to multiple questions depends on the number of points in the Likert-type scale.