|
Abstract Meta-Analysis, as a quantitative literature review method, is a decent tool to overcome reviewer-biases in traditional literature review through revising the systematic artifacts within existing researches to obtain a more accurate research conclusion than any other single study. In recent years, meta-analysis has been more and more popular among researchers of management studies. Using a sample of 79 literatures employing meta-analysis, drawn from the top management journals both at home and abroad spanning from 1980 to 2011, this study reveals two obvious trends in management research: the rapid increase in the quantity of meta-analysis literature and the fast expansion of its application fields. However, there still exist some problems while meta-analysis is applied, such as unstandardized research procedures and incongruent index selection methods, which would influence the quality of conclusions. In order to enhance the validity and reliability, this paper proposes scientifically strict research procedures and implementation steps, as well as the systematic analysis of the specific index calculation methods, index selecting standard and application conditions for each step. Based on the analysis of 79 papers, this study summarizes three objectives of meta-analysis in management research: calculate and estimate the size of the main effect and its significance; implement the homogeneity test for samples from various sources; examine possible moderating effect. At the same time, this study puts forth a seven-step procedure of meta-analysis: research problem and hypothesis determining, data preparation, mean effect size calculation, systematic artifacts correcting and hypothesis testing, moderating effect analysis, bias publication, discussion and conclusion. Moreover, this paper further emphasizes two key steps of meta-analysis: main effect analyzing and moderating effect analyzing. When applying main effect analysis, researchers must carefully revise the statistic artifacts, unify research methods, and choose proper statistic indexes. In moderating effect analysis, researchers should at first forecast the existence of moderating effect based on theories and develop related hypothesis, and then they should judge the homogeneity with multiple approaches, and describe the analysis procedures, methods, and indexes while estimating mean effect size in detail. Researchers also need to clarify the exact meaning, applicable scope, and strength and weakness of the different assessment approaches they used in moderating effect estimation. Finally, in order to contribute to the theory construction and application in this field, researchers should analyze their conclusions from theoretical perspective and point out possible directions for future research.
|
|
|
|
|
|
2012, Vol. 42 
