On November 24, 2023, we are excited to present an expert talk and a workshop on the subject of meta-analyis. Additionally, for members of the Doctoral Program, there will be an opportunity for a networking lunch, providing a platform to foster connections and delve deeper into discussions related to the topic.
How can meta-analytical methods assist us in navigating the flood of information in academic literature, underpowered individual studies, and contradictory or non-reproducible study results? This presentation introduces various meta-analytical methods and their implementation in R. It covers different types of meta-analyses and effect sizes (e.g., for correlational data, experimental studies, and longitudinal studies with mean comparisons), aiming to give listeners a conceptual understanding of which meta-analytical methods might suit their research topics and questions. Advice is provided on where and how to start a meta-analytical research project, how to conduct a systematic literature review and coding, and what challenges to be aware of. Additionally, specific R packages, such as "metafor" and "robumeta", are introduced with sample code to demonstrate their applications. The session concludes with shared "life hacks" related to meta-analyses, a discussion on typical reviewer comments on meta-analyses, and a look at the potential future directions of meta-analytic research.
Meta-analytic structural equation modeling (MASEM) refers to fitting structural equation models (such as path models or factor models) to meta-analytic data. The meta-analytic data generally consists of correlations across the variables in the path or factor model, obtained from multiple primary studies. The objective of this workshop is to learn the basics of MASEM and to get practical experience with fitting MASEM models using the dedicated online app webMASEM.
I will first contrast univariate MASEM to multivariate MASEM. Univariate MASEM refers to performing multiple univariate meta-analyses in order to obtain a synthesized correlation matrix as input in a SEM program. Multivariate MASEM in contrast involves using multivariate meta-analysis to synthesize correlation matrices across studies (e.g., GLS, TSSEM, one-stage MASEM). I will show that although univariate MASEM is the default MASEM method in for example organizational psychology, results obtained from univariate MASEM cannot be trusted.
The reason that univariate MASEM is still often used, may be that fitting MASEMs may be challenging for researchers that are not accustomed to working with R software and packages. Therefore, we developed webMASEM; a web application for MASEM. This app implements the one-stage MASEM approach, and allows users to apply multivariate MASEM in a user-friendly way. there are no prerequisites for this workshop.
The course will take place in Room B -102, Uni S, Schanzeneckstrasse 1, 3012 Bern.
If you would like to register, please fill out the Qualtrics form. The registration deadline is the 31st of October, 2023.