USE OF STATISTICS IN RECENTLY-PUBLISHED PHYSICAL EDUCATION RESEARCH (CONTENT ANALYSIS

The types and frequencies of statistical techniques reported in recently-published physical education research were studied. Also investigated were: (a) complexity of the data that were analyzed, (b) frequency and levels of significance testing, assumption testing, and data transformation, and (c) characteristics of the reporting of statistical analyses. Stratified random sampling with proportional allocation farom seven physical education research journals was used to identify a sample of 233 quantitative research reports.;Content analyses revealed that a wide variety of statistical techniques were employed in the sample of reports. Descriptive statistics were reported most frequently; a majority of the reports did, however, employ at least one inferential analysis. A p value of .05 was the most commonly reported alpha level for significance testing, although most studies failed to state a criterion alpha level. Among the inferential studies, 98% reported statistical "significance" of their findings.;The research investigations were found to be complex in terms of the number of variables studied, but less complex when one considers the number of variables simultaneously analyzed and the sample sizes employed. Multivariate analyses were employed in 25.8% of the research reports; it was suggested, however, that multivariate and repeated measures analyses should have been used more frequently than they were. Researcher writers generally did not provide readers with "help" in understanding statistics. The data analysis revealed limited observations of justifications for selection of a particular statistical technique or citations of statistical references. Very seldom did writers identify the data analysis program that was utilized.;The types and frequencies of statistical techniques employed were analyzed according to the subspecialty focus of the research. Multivariate techniques were reported most frequently in Measurement & Evaluation research and least frequently in investigations of the Functional Effects of Physical Activity. Management Theory & Practice researcher employed nonparametric techniques more frequently than did other subspecialty investigators; no nonparametric analyses were observed in research classified as Functional Effects of Physical Activity and/or Mechanical & Muscular Analysis of Motor Skills.;Findings were discussed in relation to pertinent sources cited in the review of literature. Additionally, suggestions were made for improving the quality of published physical education research and for academic preparation in statistics.