Graduation Semester and Year




Document Type


Degree Name

Doctor of Philosophy in Psychology



First Advisor

Daniel S Levine


This research examined the impact that two qualities of scales--the number of response options and the response scale label format--have on reliability and validity. Based on a meta-analytic pilot study, I expected that these two scale qualities would interact to predict reliability, such that the association between response scale length and reliability would be stronger for fully labeled than for endpoint labeled scales. This study also examined three quantitative variables that have previously been hypothesized to moderate the association between response scale length and reliability: score variability, item homogeneity, and skewness. I randomly assigned 893 participants to one of six scale format conditions to fill out six questionnaires; the reliabilities of these measures were examined. Then, the subjects took two more questionnaires, and the scores of these measures were correlated with the scores from the first six questionnaires for construct validity coefficients. Response scale length and label format did interact to predict reliability but not as expected. However, this outcome may have been due to the characteristics of the chosen sample. I found that college educated respondents had higher reliabilities at seven-point scales, as compared to five-point scales, but this pattern was not seen in the less educated group. The three quantitative variables also moderated the response scale length and reliability relationship, though not all in the manner anticipated. Finally, the number of response options did influence validity, but the only generalizable conclusion was that fully labeled scales outperformed endpoint labeled scales at seven response options.


Psychology | Social and Behavioral Sciences


Degree granted by The University of Texas at Arlington

Included in

Psychology Commons