| Abstract: |
An exploratory reliability and validity study was conducted of a relatively new response scale developed in the marketing field. Unlike many Likert-type scales, the "unbounded write-in" scale is claimed to produce distributions that more closely approximate normal distributions. This type of scale has been used in large-scale marketing studies. This experimental study sought to determine whether the scale demonstrates item test-retest reliability and whether respondents use the range of the scale in similar ways. (whether two respondents who reported the same level really felt the same way, and whether they used the same point on the scale). Participants were 220 undergraduates who responded to 10 attitudinal items taken from the 1998 General Social Survey and presented as unbounded write-in scales. Results from the study are mixed. Findings suggest that the unbounded write-in scale may offer a reliable alternative to the Likert-type scale, although the claimed advantages of its distributional qualities were not seen in this study. Focus group comments, however, suggest that the scale might not reflect similar attitudes across individuals. Further study appears warranted. Three questionnaire versions are attached. (Contains 3 figures, 4 tables, and 19 references.) (SLD) |