Analyses And Countermeasures On Statistic Application In Track-and-Field Scientific Research Dissertations

With the development of sports activities,statistic methods are increasingly useful in sports scientific and research dissertations.But through literature,some problems are found when statistic methods are applied in sports journals.So to certain problems,some feedback should be given to the sports scientific and research members.It is rather important nowadays when quantification is attached great importance by them.According to the research aim and requirement,this article adopts such research methods as literature,expert interview,and statistics. Based on three-round careful analysis and using the result of statistics, some conclusions are drawn from some selected scientific and research dissertations on track and field.1.Track-and-field researchers master and apply mostly such statistic methods as average,standard deviation,percentage,charts,t- test,one way anova,linear regression,multiple linear regression,step regression and factor analysis," etc.The total use of the above methods is up to 90%.2.In terms of experimental planning,there is such an obvious problem as neglecting group balance when dividing experimental groups.3.There are such major problems as:"average±stdev" describing skewness data;multi-group data standard deviation equality and confusing "rate" with "ratio",etc.when describing statistics.4.When referential statistics application,the big problems are parametric test instead of non-parametric test,not considering homogeneity of variance,confusing "t- test" with "one way anova", etc. 5.When conducting qualitative statistic analyses,the big problems are general chi-square test instead of accurate chi-square test,chi-square of square-chart test equation to test crosstabs file.6.When conducting relevant analyses,the main problems are linear relevant analyses on non-linear variance,order arrangement being wrong when conducting compute equation of rank correlation.7.On multiple statistics method,there are such major problems as: regression coefficient of multiple linear regression equation has not been tested;KMO test and Bartlett test have not been tested when conducting factor analysis;deciding coefficient is quite small; multiple co-linear exists in regression model,etc.8.As to explaining and illustrating statistic results,the main problems emerge as the following:not mentioning any information of statistics data,drawing conclusion only based on "P＞0.05或P＜0.05".no statistics analysis,drawing conclusion only based on superficial statistic data.The following countermeasures and suggestions are established on the basis of the drawn conclusions and the results of expert interview.1.Subjectively,pay great attention to the real role of statistics in scientific research.2.Learn and master the common-used statistic methods from others' misdoing in statistic application.3.Be careful about others' scientific research activities and statistic application in dissertations.