Métodos estadísticos utilizados en las ciencias del movimiento humano

Abstract

The purpose of this study is to describe the most cited statistical methods reported in a scientific journal of the human movement sciences area. The journal was indexed in the Science Citation Index, and was rated in 1999 with an impact factor = 2.110. Most of the 305 articles published in 2001 belonged to United States researchers (45,2%), with a clear concentration in exercise and/or sports physiology (65,9%). The most common statistical analysis performed (72%) were the following parametric techniques: analysis of variance and its different models (57,7%), student t-tests (34,8%), correlation (29,5%), and regression analysis (21%). Non parametric techniques (12,8%) reported were: chi squared (7,9%) and Wolcoxon (4,9%). Also, 12,8% of the studies showed a combined use of parametric and non parametric techniques in their data analysis. Follow-up tests (post hoc), as well as other assumption analysis were reported. Researchers (61,6%) did not report using a statistical or computerized package to analyze their data. From those who did use a package (38,4%), 13,4% preferred to use the Statistical Package for the Social Sciences to analyze their data. In summary, most studies published in a high-quality journal were properly designed and data analyzed, relying on advanced statistical analysis. Physical educators and movement sciences specialists are suggested to modify their research methods and statistical analysis courses in order to provide a better understanding of the statistical techniques reported in this document. Key words: research methods, bibliometrics, statistics, human movement sciences.