23551c
Seminar am PC
WiSe 25/26: S-PC Introduction to Structural Equation Modeling with Linear, General Linaer and Mixed Models in R
Oksana Buzhdygan, Felix May
Hinweise für Studierende
Additional module information: Introduction to Structrual Equation Modeling
Schließen
Zusätzl. Angaben / Voraussetzungen
Prior knowledge in R and linear models including regression, ANOVA and ANCOVA is required. Please use the computer not a tablet because R is difficult to install on a tablet.
Kommentar
Content:
Seminar on the PC:
In the seminars on the PC, students practically apply the topics and methods, learned during the lectures and seminars. Using a number of worked examples from the published ecological literature, students develop, evaluate, modify and solve the Structural Equation Models (SEM) for different data types (using linear, generalized linear and mixed effect models) using the R software under supervision and later independently. Students practice the selection of data analysis strategies using SEM for different datasets (e.g., random vs. not random samples; count vs proportional or binary data; numeric vs categorical predictors). With SEM models students analyse cause-effect connections, test direct and indirect effects and interpret the mechanisms in the study systems. With piecewise SEM students analyse grouped and nested data, calculate marginal means for categorical predictors, test the interactive effects of predictors using SEM and interpret the results in the ecologically meaningful contexts.
Learning objectives:
In this module the students acquire the following knowledge and skills:
Seminar on the PC:
In the seminars on the PC, students practically apply the topics and methods, learned during the lectures and seminars. Using a number of worked examples from the published ecological literature, students develop, evaluate, modify and solve the Structural Equation Models (SEM) for different data types (using linear, generalized linear and mixed effect models) using the R software under supervision and later independently. Students practice the selection of data analysis strategies using SEM for different datasets (e.g., random vs. not random samples; count vs proportional or binary data; numeric vs categorical predictors). With SEM models students analyse cause-effect connections, test direct and indirect effects and interpret the mechanisms in the study systems. With piecewise SEM students analyse grouped and nested data, calculate marginal means for categorical predictors, test the interactive effects of predictors using SEM and interpret the results in the ecologically meaningful contexts.
Learning objectives:
In this module the students acquire the following knowledge and skills:
- Gain basic knowledge of structural equation modeling (SEM) framework and path analysis
- Learn how to develop, evaluate, refine, solve, and interpret structural equation models
- Master basic skills to analyze data with SEM in the software R
- Gain basic knowledge of piecewise SEM and how it differs from the classical SEM
- Master basic skills to implement in the SEM count, binary, proportion, and categorical response variables, as well as nested data with the mixed effect models and piecewise SEM using the software R
- Gain basic understanding of causal relations, bottom-up and top-down control, and how to calculate direct and indirect effects in ecological and biological systems (e.g., communities, food webs, ecosystems)
- Independently apply SEM for different data types
- Present statistical methods and results in oral and written form to a specialist audience.
Literaturhinweise
Grace (2006) Structural Equation Modeling and Natural Systems. Cambridge Univ. Press.
Shipley, B. (2016). Cause and correlation in biology: A user's guide to path analysis, structural equations and causal inference with R. Cambridge university press.
Lefcheck (2021) Piecewise Structural Equation Modeling in Ecological Research: https://jslefche.github.io/sem_book
Shipley, B. (2009). Confirmatory path analysis in a generalized multilevel context. Ecology, 90(2), 363-368.
Schließen