体育科学研究中有中介的调节和有调节的中介潜变量模型检定

Methods for Testing Moderated Mediation and Mediated Moderation Latent Variable Models in Sports Science Research

体育科学研究中,中介变量和调节变量2个第三变量同时出现的检定情形越来越多。有调节的中介和有中介的调节模型检定,关注自变量与因变量关系的机制和边界条件,极大促进了体育科学研究的深入发展。随着统计技术和软件开发的长足进步,则更有裨益。就体育科学研究中,存在的有调节的中介和有中介的调节潜变量模型检定问题进行整理和分析,并参照目前检定力较强、学界一致推崇的偏差校正非参数百分位Bootstrapping法(Bias-corrected non-parametric percentile Bootstrap method)进行案例检定分析。结论:(1)体育科学研究中,有中介的调节和有调节的中介潜变量检定存在的主要问题是变量关系利用不充分、理解及表述不规范、调节变量拉入时机缺乏理论支撑、采用标准化值检定、检定方法失当等问题;(2)Bootstrapping法的实用性、易用性和可读性,有利于满足体育科学研究中自变量与因变量关系机制和边界条件的检定需求;(3)有调节的中介和有中介的调节潜变量模型的检定程序和方法包括检定中介效果跨群组特性、检定中介效果改变的强度与方向和检定中介效果的差异(异质性检定)3个步骤。

In sports science research,the simultaneous appearance of two third variables,mediating variables and moderator variables,has become increasingly common. Moderated mediation and mediated moderation models verification focuses on the mechanism and boundary conditions of the relationship between in-dependent variables and dependent variables,which greatly promotes the in-depth development of sports science research. With the rapid progress in statisticaltechnology and software development,it will be even more beneficial. The study has collated and analyzed the verification problems of moderated mediation andmediated moderation latent variable models in sports science research. Case verification analysis was conducted with reference to the Bias-corrected non-parametric percentile Bootstrap method,which is currently highly respected by the academic community. The main problems of verification of moderated media-tion and mediated moderation latent variable models in the study of sports science are the insufficient use of variable relationships,the lack of understandingand expression,the lack of theoretical support for the timing of adjustment variables,the use of standardized value verification,and improper verification meth-ods. The practicality,ease of use and readability of the Bootstrapping method are conducive to meeting the verification requirements for the relationship mecha-nism and boundary conditions in sports science research. There are three steps in the verification procedures and methods:the effectiveness of cross-group verification of mediating effects,the strength and direction of changes in mediation effects,and the differences in assay mediation effects(heterogeneity).