Reliability and construct validity of Turkish version of physical education activities scale

Abstract

Uğur Altay Memiş

This research was conducted to examine the reliability and construct validity of Turkish version of physical education activities scale (PEAS) which was developed by Thomason (2008). Participants in this study included 313 secondary and high school students from 7th to 11th grades. To analyse the data, confirmatory factor analysis, post hoc analysis and modification indices were conducted by AMOS Software 21.0 and internal consistency for reliability was tested by SPSS 13.0. Findings showed that GFI (.885), CFI (.923), RMR (.227), and RMSEA (.055) of four factor correlation model was best for representing Turkish version of PEAS factorial structure. Cronbach’s alpha coefficients of the subscales (behavioral control: .74; social-norms: .75; physical education enjoyment: .85; self-efficacy: .50) and the Cronbach’s alpha coefficient of the total scale (.90) proved that Turkish Version of Physical Education Activity Scale had internal consistency. As a result, Turkish version of PEAS including four subscales and 24 items was reliable and had construct validity to measure 13 to 18 years old adolescents’ perceptions of environmental factors that influence their behavior directly or indirectly when they participate in physical activities in physical education classes.