Statistical methods of causal inference in sports science research

Abstract

Among the complex large quantity of sports phenomenon and sports issues, we need to discuss the causal relationship and essential rule in between. Mathematical statistics is one of the common and practical methods for causal inference. In order to better understand, select, and use mathematical statistic methods, and to avoid deviation of research result due to misuse or even abuse of statistic methods, this paper, through literature and comparative analysis, enumerates the basic theory, model of calculation formula, applicable condition, and operation process of common causal relationship statistical methods, including partial correlation analysis, multiple regression analysis, path analysis, and structural method model, with raising limitation of these methods at the same time. These methods have the same logic, but different principles and calculation process, with one of them a panacea. Therefore, research result of high reliability and validity can only be drawn by full understanding of the applicable conditions of the various methods and comprehensive use of various methods based on actual condition.