Background/Purpose In multiple regression analysis, most researchers judge the relative contribution of each of the predictor variables based on their respective beta weights (Coorley & Lohnes, 1971). However, it is argued that the interpretation of results based solely on beta weights can lead to erroneous conclusions regarding the importance or relevance of predictor variables. In general, both beta weights and structure coefficients must be interpreted in order to reach valid conclusions (Bowling, 1993). The purpose of this study is to investigate the use and the interpretation of structure coefficients of the published articles in the Journal of Teaching in Physical Education.
Method The researcher reviewed all 129 research articles published in the Journal of Teaching in Physical Education from Volume 24, Issue 1 to Volume 29, Issue 2. Articles with multiple regression analysis were examined for the reporting of regression coefficients and structure coefficients. The researcher also examined how the interpretations might have differed if standardized regression coefficients and structure coefficients had been interpreted.
Analysis/Results Of 19 regression analyses performed, only one article reported and interpreted structure coefficients. Two examples were found on how misinterpretations or incomplete interpretations were made because the researchers relied on beta weights when interpreting multiple regression results.
Conclusions Both beta weights and structure coefficients ought to be interpreted when noteworthy regression results are isolated (Courville, 2001; Thompson, 1997, 2006). Relying only on the interpretation of beta weights without also considering the structure coefficients may result in researchers unconsciously misinterpreting and choosing incorrect predictor variables.