Meta-Analysis: Processes and Prospects

Article excerpt

The proliferation of research in recent years has expanded the need for more careful, accurate, and complete reviews in the exercise and sport sciences. Meta-analysis, a quantitative statistical approach to reviewing the literature, offers a viable, generally less subjective, alternative to the traditional qualitative review. However, undertaking this process requires some expertise. Therefore, the purpose of this symposium is to generally introduce the process while providing detail on important aspects of meta-analytic reviews: searching the literature, coding the data, and statistically analyzing effect size data. The symposium will conclude by examining past and present uses of this technique with ideas for future application.

An Introduction to Meta-Analysis

V. Gregory Payne, San Jose State University

The need for scientific reviews of primary research has expanded with the vast proliferation of studies conducted in recent decades. Meta-analysis is a technique that has been designed to quantitatively "sum-up" a body of literature on a selected, specific research problem using formal statistical techniques. If properly conducted, meta-analysis holds a major advantage over traditional qualitative reviews because conclusions can be more objectively formulated based on the quantitative results. Many believe this process is a "wave of the future" with powerful prospects. Previous meta-analyses have led to exciting, sometimes controversial, conclusions with potential to overturn previously ingrained scientific tradition. Examples will be provided and discussed. The purpose of this paper is to generally examine the process of meta-analysis. Topics will include how to select a viable problem, tips on searching the literature, calculation of the effect size, statistical analysis, and interpretation of results. All topics will be covered generally as an introduction and precursor to the more specific presentations that follow. Specific examples, based on past meta-analyses conducted by the authors and others, will also be provided.

Literature Search and Data Coding for Meta-Analysis

Jin H. Yan, The University of Houston

Searching literature and coding data are two important parts of meta-analytic research. The purpose of this presentation is to introduce the basic procedures of locating specific research articles and transforming the information into effect size (ES) for quantitative analysis. Studies for a meta-analytic review can be obtained many ways. This includes reference lists of previously published reviews, textbooks, previously published qualitative reviews, and searching computer databases (MEDLINE, Psych-LIT, Psych-INFO, ERIC) using specific key terms. In addition, a manual searching of the references of related published studies should be performed during the review process. Once the final list of studies to be reviewed has been assembled, researchers peruse and carefully study each paper to determine those that meet predetermined criteria for use in later stages of the meta-analysis. Papers that meet these criteria are coded. The objective of the coding process is to translate information from each individual p aper into a precise, concise integrated form of data storage. Recorded information might include general characteristics of the study (e. …