Analysis of Likert-scale data: A reinterpretation of Gregoire and Driver.

Abstract

Most studies that have investigated the use of coarsely grained scales have indicated that the accuracy of statistics calculated on such scales is not compromised as long as the scales have about 5 or more points. Gregoire and Driver (1987), however; found serious perturbances of the Type I and Type II error rates using a 5-point scale. They carried out three computer simulation experiments in which continuous data were transformed to Likert-scale values. Two of the three experiments are shown to be flawed because the authors incorrectly specified the population mean in their simulation. This article corrects the flaw and demonstrates that the Type I and Type II error rates are not seriously compromised by the use of ordinal-scale data. Furthermore, Gregoire and Driver's results are reinterpreted to show that in most cases, the parametric test of location equality shows a power superiority to the nonparametric tests. Only in their most nonnormal simulation does a nonparametric test show a power superiority.

Students' perceptions of assessment in undergraduate mathematics

“…While there is much discussion about the analysis of Likert scales with parametric methods, we note that Rasmussen (1989) demonstrates that they result in increased type I or II error rates only in the most non-normal of situations.…”

Assessing the Psychometric Properties of the Attitudes Toward Seeking Professional Psychological Help Scale-Short Form in Mainland China

“…First, the measure was initially constructed with a 4-point rating scale (strongly disagree, disagree, agree, strongly agree), which may be problematic, as narrow ranges of response options tend to be associated with lower reliability and higher susceptibility to Type II errors (Rasmussen 1989). Previous studies have demonstrated that instruments employing rating options ranging from 1 to 5 tend to be less reliable than those with scales from 1 to 7 (Krosnick and Fabrigar 1997;Oaster 1989).…”

Examining private forest policy preferences

“…While these Likert scale variables are ordinal, ordinal data of five or more classes are generally treated as continuous and analyzed using statistical analyses such as regression, analysis of variance, and correlation (Borgatta and Bohrnstedt, 1980). Although there has been some debate regarding the use of Likert scales as continuous (Wilson, 1971), many argue that such procedures are reliable and robust (Binder, 1984;Davison and Sharma, 1988;Davison and Sharma, 1990;Gregoire and Driver, 1987;Rasmussen, 1989;Zumbo and Zimmerman, 1993). In a recent review of the literature on this topic, Jaccard and Wan (1996) concluded that the use of ordinal scales in statistical procedures assuming interval data did not compromise Type I and Type II error.…”