Meta-analysis, a statistical method to summarize independent studies under the same subject, is considered as the best quantitative summary method up to date. Effect size is its metric. Cumulative effect size and total heterogeneity,the most important parameters in meta-analysis, are the index which denote the center trend and variance of effect sizes in meta-analysis respectively. The method of their calculation depends on the different structure data. In this paper, three calculation methods suited to three different structure data (no structure data, categorical structure data and continuous structure data) were introduced and compared.
[2] Cooper H M. 1998. Synthesizing research: a guide for literature review. (3rd ed.) [M]. Thousand Oaks: Sage.88-95.
[3] Hedges Ⅴ, Olkins Ⅰ. 1985. Statistical Methods for Meta-Analysis [M]. New York: Academic Press. 55-59.
[4] Gurevitch J, Hedges Ⅴ. 1993. Meta-analysis combining the results of independent experiments[M]. New York: Chapman & Hall. 114-119.
[5] Williams P A, Haertel E H. 1982. The Impact of leisure time television on school learning: A research synthesis[J].American Educational Research Journal. 19(41): 19-50.
[7] Goldberg D E, Rajaniemi T. 1999. Empirical approaches to quantifying interaction intensity: Competition and facilitation along productivity gradients[J]. Ecology,80:1118-1131