AI Summary
[DOCUMENT_TYPE: instructional_content]
**What This Document Is**
These are lecture notes from EVAL 6970, a graduate-level course on Research on Evaluation at Western Michigan University. The notes specifically focus on advanced meta-analytic techniques, moving beyond basic meta-analysis to explore methods for investigating the factors that explain differences in study results. Key topics covered include meta-regression – a statistical approach for examining how study characteristics relate to effect sizes – and the handling of complex data structures within meta-analysis. The material appears to be based on a lecture delivered in Spring 2011.
**Why This Document Matters**
This resource is invaluable for graduate students and researchers in fields like education, psychology, public health, and social work who utilize or plan to utilize meta-analysis in their research. It’s particularly helpful for those seeking to understand how to account for variability *between* studies, rather than simply calculating an overall average effect. Anyone preparing to conduct a meta-analysis involving multiple potential moderator variables, or working with datasets that have non-independent observations, will find this a useful reference. It’s best used as a companion to a core meta-analysis course or textbook.
**Common Limitations or Challenges**
These notes represent a specific lecture and do not constitute a comprehensive textbook on meta-analysis. They do not provide step-by-step instructions for performing these analyses in statistical software. The notes also assume a foundational understanding of basic statistical concepts, including regression and ANOVA. While examples are referenced, the specific datasets and results used in illustration are not fully detailed within this preview.
**What This Document Provides**
* An overview of the purpose and application of meta-regression.
* Discussion of recommended ratios for the number of studies to covariates in meta-regression.
* Presentation of fixed-effect model outputs, including regression tables and ANOVA information.
* Exploration of random-effects models and their application to meta-regression.
* Analysis of variance components within and between studies.
* Discussion of the proportion of variance explained by covariates.
* Reference to an in-class activity utilizing specific data for practical application.