Designed to provide researchers clear and informative insight into techniques of meta-analysis, the Third Edition of Methods of Meta-Analysis: Correcting Error and Bias in Research Findings is the most comprehensive text on meta-analysis available today. It is the only book that presents a full and usable treatment of the role of study artifacts in distorting study results, as well as methods for correcting results for such biases and errors.
Chapter 3: Meta-Analysis of Correlations Corrected Individually for Artifacts
Meta-Analysis of Correlations Corrected Individually for Artifacts
Introduction and Overview
In Chapter 2, we examined 11 study design artifacts that can affect the size of the correlation coefficient. At the level of meta-analysis, it is possible to correct for all but one of those artifacts: reporting or transcriptional error. Except for outlier analysis, we know of no way to correct for data errors. Outlier analysis can detect some but not all bad data and is often problematic in meta-analysis (as discussed in Chapter 5). Sampling error can be corrected for, but the accuracy of the correction depends on the total sample size that the meta-analysis is based on. Our discussion of meta-analysis in this chapter and in Chapter 4 implicitly assumes that the meta-analysis is ...