The following training videos are designed for experienced systematic reviewers and present advanced methods for conducting systematic review.
These videos come from the Center for Evidence-Based Crime Policy-Campbell Collaboration Joint Symposium on Evidence-Based Policy, held August 15 – 16, 2011 at George Mason University. The training workshops were part of a larger event aimed at engaging a wide range of people in the ongoing discussion on how to develop, use and improve evaluation research for public policy decision makers.
Who are these training videos meant for?
- Researchers with experience in carrying out a systematic review
- Policy makers interested in learning more about systematic reviews and evidence based policy
- Evaluation researchers
Calculating effect sizes, advanced
This training video guides participants through computing standardized mean difference, odds ratios, and correlation effect sizes from information reported in study reports and emphasizes the dilemmas of calculating effect sizes when primary studies do not provide needed statistics.
Presented by David Wilson – Professor & Chair, Department of Criminology, Law & Society, George Mason University; Editor, Campbell Crime & Justice Coordinating Group.
Presentation file
Fixed vs. random effects models
This training video provides paraticipants with an overview of the differences between fixed and random effects analysis models in meta-analysis, and the issues reviewers should consider when making the choice between these models. Participants can go through computational examples comparing fixed and random effects models to see how the two models differ in the types of statistical inferences that they support.
Presented by Terri Pigott – Professor of Research Methodology, Loyola University; Editor, Campbell Methods Group
Presentation file
Cluster adjustments in computing effect sizes
This training video shows participants how to calculate effect size statistics when primary studies use cluster randomized designs (e.g., assigning schools or teachers to an intervention), and the analytic issues that arise when including cluster level studies. Topics include an overview of when cluster adjustments to effect size statistics are necessary, analysis methods for cluster adjustments, and how to deal with research literatures that have both individual level and cluster level effect sizes.
Presented by Terri Pigott – Professor of Research Methodology, Loyola University; Editor, Campbell Methods Group.
Presentation file
Using robust standard errors for dependent effect sizes
This training video provides an overview of the use of robust standard errors for analyzing dependent effect size estimates. The most common form of dependence occurs when studies provide multiple measures of the same underlying construct (e.g., self-reports and parent-reports of child behavior). The training goes over when robust standard errors are needed and points out the available software for using this approach. Participants are then guided through a demonstration of how to analyze and interpret results from meta-regression models using robust standard errors.
Presented by – Elizabeth Tipton, Assistant Professor of Applied Statistics, Teachers College, Columbia University; Emily Tanner-Smith, Research Associate, Peabody Research Institute, Vanderbilt University; Associate Editor, Campbell Methods Coordinating Group & Mark Lipsey, Director, Peabody Research Institute, Research Professor of Public Policy, Vanderbilt University; Campbell Collaboration Co-Chair, Steering Committee
Presentation file