文章基本信息

标题：Deciding on the Scale Granularity of Response Categories of Likert type Scales: The Case of a 21‑Point Scale
本地全文：下载
作者：Noel Pearse
期刊名称：Electronic Journal of Business Research Methods
印刷版ISSN：1477-7029
电子版ISSN：1477-7029
出版年度：2011
卷号：9
期号：2
页码：159-171
出版社：Academic Conferences International Limited
摘要：This research investigates the use of a 21‑point Likert type scale in the design of a questionnaire that explores factors related to staff turnover and retention. The paper examines the notion of granularity in researcher‑defined fixed rating scales, where granularity refers to the number of response categories or cut off points that are imposed on a scale (Smithson 2006). The aim of this research paper is to examine the usefulness of a scale with high granularity, from the perspectives of respondents and the researcher. The questionnaire was administered among employees in three different public sector organisations in South Africa, to create a combined data set of 178 respondents. Informing the formulation of the hypotheses was Parducci’s (1965 cited in Tourangeau, Rips & Rasinski 2000) range‑frequency model, which assumes that respondents make use of the various response categories available with equal frequency, if they are evenly spaced. It was therefore hypothesised that (1) there are no significant differences in the frequency of use of the 21 response categories, implying that all of the response categories are useful to respondents; (2) that there would be no difference in the response pattern of respondents when different scale types and lengths are used, implying that increasing the scale granularity did not lead to redundancy; and (3) that there are no significant differences in the variation of responses with ongoing use of the scale. That is, if the scale was useful to respondents, they would continue to use a wide range of the response options available. Chi‑square tests were primarily used to test the hypotheses. It was concluded that the 21‑point scale was useful to respondents and by implication to researchers as well. This was evident in the spread of responses across the 21 response categories of the scale, and that even with prolonged use, they continued to utilise a wide range of response options. It was recommended that researchers should give more explicit attention to scale granularity when designing a questionnaire and that further research is required to assess the value of various levels of scale granularity.
关键词：questionnaire design; scale construction; likert scale; scale granularity