Evidence synthesis includes systematic reviews, meta-analyses, and other forms of evidence such as systematic maps or research weaving that allow us to summarize knowledge. When preceded by a systematic review, meta-analysis is a powerful statistical tool for quantitatively integrating findings across studies to (1) test overall effects and their generalizability, (2) understand context-dependencies, and (3) generate and test second-order hypotheses. Systematic reviews and meta-analyses have become a standard approach for qualitatively and quantitatively synthesizing evidence across fields, and their use has increased exponentially in the last decade. This course will provide a hands-on overview and introduction to modern methods for evidence synthesis, with a special focus on systematic review and meta-analysis in ecology and evolution.
We will begin with a detailed overview of the systematic review approach, focusing on question formation, systematic searching and study screening. Next, we will focus on the meta-analytic process, specifically on effect size choice, data extraction, data analysis, and importantly, result interpretation. We will make use of multilevel meta-analysis and meta-regression, with an introduction on how to account for phylogeny when multiple species are synthesised. Since meta-analytic results cannot be interpreted without a deep understanding of heterogeneity and publication bias, we will cover in detail how to estimate, adjust for, and interpret heterogeneity and publication bias.
This course will include a mix of lectures and hands-on exercises using real meta-analytic datasets. The emphasis throughout the course is on the application of the various methods and the interpretation of the results using the free software R and the R packages ‘metafor’ (Viechtbauer 2010) and ‘orchaRd’ (Nakagawa et al. 2023). The course will follow the principles of open science, with a strong focus on the importance of adhering to preferred reporting items for systematic reviews and meta-analyses in ecology and evolutionary biology (PRISMA EcoEvo; O’Dea et al. 2021). Throughout, we will consider examples of how to interpret results and present them using tables and data visualization, and for each step, we will provide literature and practical resources (e.g., R scripts).

 This course is designed to be as self-contained as possible. However, we will assume basic familiarity with research design and statistical concepts (e.g., hypothesis testing, variance, correlations, linear regression). While it is not strictly necessary, it would help the participants to familiarise themselves with systematic reviews and meta-analyses before the course, either by reading brief introductions (see references below) or looking at their favourite examples from their field. Although we will present some formulas and statistical notations, the course will use conceptual understanding and visualisation to understand and interpret meta-analytical models.  
All the computations will be done using R. We will go through all the basic (and more advanced) steps of conducting systematic reviews, meta-analyses and meta-regressions during the course and all the R scripts will be provided. Although this is not an R programming course, some familiarity with R is necessary. If you are new to R, please explore an introductory resource ahead of time (here is one you could try).

The practical components of this course require the free software R (and R-Studio). Before the course, please install the latest version of R from the Comprehensive R Archive Network (CRAN) for your operating system – Windows, MacOS or Linux. If you already have R installed, please check that it is the current and if not, please update it. Although not strictly necessary, we will run R from within another free software, RStudio. RStudio is a powerful and rather popular user interface for R.

Please, download and install (or update) RStudio from posit.
Once R and RStudio are installed, install (and make sure they work) the following packages by running this code in R:
install.packages("pacman") #it makes loading, installing, and updating R packages super easy!
pacman::p_load(tidyverse,metafor,revtools,synthesisr,metaDigitise,ape,rotl,ggpubr,orchaRd)

References
Foo et al. 2021. A practical guide to question formation, systematic searching and study screening for literature reviews in ecology and evolution. Methods in Ecology and Evolution 12 (9): 1705-1720. https://doi.org/10.1111/2041-210X.13654

Gurevitch et al. 2018. Meta-analysis and the science of research synthesis. Nature 555: 175–182. https://doi.org/10.1038/nature25753

Nakagawa et al. 2023. orchaRd 2.0: An R package for visualizing meta-analyses with orchard plots. Methods in Ecology and Evolution 14 (8): 2003-2010. https://doi.org/10.1111/2041-210X.14152

O’Dea et al. 2021. Preferred reporting items for systematic reviews and meta-analyses in ecology and evolutionary biology: a PRISMA extension. Methods in Ecology and Evolution 96 (5): 1695-1722. https://doi.org/10.1111/brv.12721  

Viechtbauer. 2010. Conducting meta-analyses in R with the metafor package. Journal of Statistical Software, 36(3): 1-48. https://doi.org/10.18637/jss.v036.i03

Viechtbauer. All about the metafor Package. https://www.metafor-project.org/doku.php/diagram

Monday – Classes from 13:00-17:00 CET
Lectures and exercises on question formation and how to perform literature searches for scientific literature, including grey literature, in search platforms such as Web of Science, Scopus and PubMed.

Tuesday – Classes from 13:00-17:00 CET
Lectures and exercises on title-and-abstract and full-text screening, including the use of decision trees to increase inter-observer agreement and reproducibility, and an introduction to Risk of Bias assessments.

Wednesday – Classes from 13:00-17:00 CET
Lectures and exercises on effect size choice, data extraction and data analysis (meta-analysis and meta-regression). The focus will be on the meta-analysis of correlations (r, Zr) and mean differences (lnRR, SMD), but we will also introduce the meta-analysis of variance (lnCVR, lnVR) and other less commonly used effect sizes in ecology and evolution.

Thursday – Classes from 13:00-17:00 CET
Lectures and exercises on data analysis (continuation), heterogeneity and publication bias with a focus on how to estimate heterogeneity using a pluralistic approach, the importance of prediction intervals, and how to test for and adjust for publication bias (e.g., small-study and decline effects) using multilevel meta-regressions.

Friday – Classes from 13:00-17:00 CET
Lectures, exercises and discussion on result interpretation, followed by a Q&A session (including about your own ongoing evidence synthesis projects, so bring them along).

Should you have any further questions, please send an email to info@physalia-courses.org