# Meta-analysis

**Overview:**

The goal of this workshop is to serve as an introduction to the use of meta-analysis to combine summary data. Specifically, we will cover the different assumptions required and types of analyses one can use when trying to synthesize information from different sources. We will begin by considering the case of estimating a mean for single characteristic from different studies. We will discuss different estimators and diagnostic plots that can be used estimate quantities of interest and to assess the assumptions necessary for proper inference. Then we will consider methods for synthesizing comparisons amongst two or more groups across different contexts and different approaches that can be used. Finally, we will briefly discuss how to use study level information to adjust for heterogeneity across data sources.

Participants will get access to several worked examples written in R, mostly using the meta and metafor R packages.

At the end of this workshop, you will be able to:

- Understand the assumptions necessary for proper synthesis of estimates from multiple data sources.

- Understand the different estimators that can be used and which estimators are likely to perform better in which context.

- Understand the differences between arm-based and contrast-based estimation for synthesizing measures from multiple groups.

- Conduct exploratory analyses, fit standard meta-analysis models and assess fidelity to model assumptions using the meta and metafor packages in R.

**Prerequisites:**

- An undergraduate/graduate introduction to statistics (basic knowledge of means, variances, hypothesis testing and confidence intervals);

- Basic knowledge of R (basic R commands and understanding of how to analyze data in R)

**Date:** Friday, 12 May 2023.**Time:** 10 a.m. to 12 p.m.**Location:** hybrid (in-person at Burnside Hall 1104, and online via Zoom).**Instructor:** Prof. Russell Steele, Dept. of Mathematics and Statistics, McGill.