Research designed to inform practice and policy decisions is based upon the reliability and validity of the measures used. Although Likert scales are perhaps the most widely used method for constructing scales, a number of little known problems are associated with their use, including multidimensionality and relatively coarse ordinal level data that taps only a portion of the underlying theoretical construct (Visser, Krosnick & Lavrakas, 2000). To circumvent the problems associated with Likert scales, phrase completion scales were developed. Phrase completion scales are concise, unidimensional measures designed to tap ordinal level data in a manner that approximates interval level data. The purpose of this study was to examine the psychometric properties of phrase completions scales relative to Likert scales. To compare the reliability and validity of the two methods of scale construction, traditional Likert items were transformed into phrase completion items. The two methods were compared using a convenience sample of graduate students (N = 134). Based upon the psychometric properties recorded preliminary testing, a sample size in excess of 100 was projected to produce stable parameter estimates (Guadagnoli & Velicer, 1988; MacCallum, Widaman, Zhang & Hong, 1999). Content analysis of the written comments suggested that respondents preferred answering questions that used the phrase completion method relative to those using the traditional Likert format. The phrase completion method was also associated with higher Cronbach's alphas and higher inter-item correlations. Similarly, the factor loadings were stronger and, correspondingly, more of the variance was explained using the phrase completion items. These results were confirmed by the validity, measurement error, and reliability coefficients obtained using SEM generated measurement equations. The validity and reliability coefficients were higher for the phrase completion items relative to the Likert items while the levels of measurement error were lower. The results suggest that scales constructed using the phrase completion method may offer a better method of measurement compared to similarly Likert items, at least in some situations. The difference may be particularly pronounced when attempting to measure attitudes that span the full length of the underlying theoretical continuum of the construct being measured. For example, items designed to tap perceptions of intervention effectiveness, in which the continuum might range from ineffective to completely effective, may be better measured using phrase completions than Likert items.

Guadagnoli, E., & Velicer, W. (1988). Relation of sample size to the stability of component patterns. Psychological Bulletin, 103, 265-275. MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor analysis. Psychological Methods, 4(1), 84-99. Visser, P. S., Krosnick, J. A., & Lavrakas, P. J. (2000). Survey research. In H. T. Reis & C. M. Judd (Eds.), Handbook of research methods in social and personality psychology (pp. 223-252). New York: Cambridge University Press.