Effect of Varied Response Alternatives and Sample Size on Reliability of an Attitude Scale Physical Activity Among Female Students in College of Basic Education

Abstract

Objectives: To examine differences in the reliability coefficients of a psychological scale according to the number of response alternatives and sample size. The study also sought to determine an appropriate number of response alternatives and sample size for estimating the reliability of psychological scales in physical education and sport sciences.

Materials and Methods: A descriptive design incorporating survey and comparative methods was employed. The participants were 150 third-year female students, representing 13.63% of the target population. The Attitudes Toward Physical Sports Activity Scale was administered in three response formats comprising three, four, and five alternatives. Reliability coefficients were calculated and compared across the response formats and different sample sizes.

Results: The reliability coefficients obtained from the three-, four-, and five-alternative formats were similar when the sample size exceeded 30 participants. As the sample size increased beyond 30, the coefficients converged regardless of the number of response alternatives used. Thus, sample size appeared to have a more meaningful influence on the stability of reliability estimates than differences among the three response formats.

Conclusions: Psychological scales using three, four, or five response alternatives can produce comparable reliability coefficients when administered to samples larger than 30 participants. Researchers in physical education and sport sciences should therefore use more than 30 participants when estimating scale reliability to obtain more stable and dependable coefficients.