Abstract
The purpose of this study was to compare the reliability, variability, and predictive validity of two common scaling response formats (semantic differential, Likert-type) and two numbers of response options (5-point, 7-point) in the physical activity domain. Constructs of the theory of planned behavior were chosen in this analysis based on its high frequency of application in exercise and physical activity. The participants were 412 undergraduate students who completed measures of the theory of planned behavior and self-reported physical activity two weeks later. One of four questionnaires, each containing a scaling response format, were distributed randomly and formed four groups of approximately n = 100 for comparisons (5-point Likert, 5-point semantic differential, 7-point Likert, 7-point semantic differential). Results showed that the 7-point options had greater variability than the 5-point options and that the 7-point Likert scale had the highest overall reliability. These differences, however, did not translate into predictive validity of behavior. The findings support the use of all of these types of scales with physical activity research because of their relatively equivalent outcomes.
ACKNOWLEDGMENTS
RER is supported by a New Investigator award from the Canadian Institutes of Health Research and with funds from the Social Science and Humanities Research Council of Canada and the Canadian Diabetes Association. DHM is supported by a Vancouver Island University Research Award. This authors also wish to acknowledge the hard work of data collection and data entry by Megan Yim.
Notes
1 Measurement of self-efficacy is sometimes measured in numerical gradation (CitationBandura, 1997). This style was not assessed.