Search
GESIS Training Courses
user_jsdisabled

Wiss. Koordination

Alisa Remizova

Administrative Koordination

Janina Götsche

Please wait...

Item Response Theory in R

About
Location:
Online via Zoom
 
Course Duration
Mo-We: 10:00-17:00 CET
 
General Topics:
Data Analysis
Course Level:
Beginner
Format:
Workshop
Software used:
R
Duration:
3 days
Language:
English
Fees:
Students: 330 €
Academics: 495 €
Commercial: 990 €
 
Keywords
Additional links
Lecturer(s): Sebastian Weirich, Nicklas Hafiz

About the lecturer - Sebastian Weirich

About the lecturer - Nicklas Hafiz

Course description

Registration deadline: 29 March 2027 
Registration after this date might be possible, but is subject to sufficient demand and availability of places.
 
Item response theory (IRT) is an indispensable tool for modeling latent constructs from test or questionnaire data. It can be used to analyze cross-sectional and longitudinal data to test the corresponding research hypotheses, and to evaluate the quality of a measurement instrument. This workshop will discuss the advantages IRT offers over classical test theory and sum scores, and when those advantages matter. You will learn the principles of latent measurement models and the main parameterizations of IRT models (the Rasch model, 2PL, 3PL, and the partial credit model). We will also cover more technical topics, such as the differences between estimation methods (joint, conditional, and marginal maximum likelihood) and plausible values, which allow group-level statistics to be estimated accurately without bias from individual-level measurement error. The goal will be to prepare you to use IRT models in your own work. Using code examples and exercises in R, we will show how to check the assumptions that must hold for IRT methods to be appropriate, for example, by testing for differential item functioning (DIF), and explore different use cases from psychological and educational research. You are welcome to bring your own data to work on during the exercises.
 
Organizational structure of the course  
  • The workshop combines theoretical lectures with practical exercises. Usually, each block lasts 90 minutes, and each workshop day consists of four blocks. The practical exercises apply the topics covered in the lectures to example data but also explore more advanced topics in a more practical fashion.
  • You can work on the exercises on your own or in small groups and help each other. The instructors are also available for support during the exercises.


    • Target group

  • The workshop is aimed at people who want to analyze test data that operationalize latent constructs with IRT models. One use case is pilot studies meant to assess the psychometric quality of the items. A typical question here would be: Are the items in the test suitable for measuring the construct „intelligence“? Are the items fair, do they discriminate well, and do they all measure the same construct?
  • The analyses can also address substantive research questions, for example: How strongly are latent traits (e.g., academic abilities) correlated with certain personal characteristics (SES, gender)? Do students who complete a particular training make greater progress in mathematics than a control group?


    • Learning objectives

    By the end of the course, you will be able to:
  • prepare data for IRT analysis;
  • conduct IRT analysis:
  • test assumptions;
  • choose a feasible model;
  • identify and exclude items that are not suitable for IRT analysis;
  • answer substantive research questions by specifying confirmatory models that allow specific statistical hypotheses to be tested;
  • generate derived data like plausible values from IRT models;
  • use the literature and the overview of relevant R packages provided in the workshop for further self-study and your own use cases.


    • Prerequisites

  • Basic R knowledge: data management; reading, creating and manipulating objects; saving objects and data sets.
  • Basic knowledge of statistics and probability theory.
    •  
     
    Software requirements
    R, free to install from:
  • Windows: https://cran.r-project.org/bin/windows/base/
  • Mac: https://cran.r-project.org/bin/macosx/
  • Linux: https://cran.r-project.org/
    • R packages:
  • remotes: install.packages("remotes")
  • lme4: install.packages("lme4")  # at least version 1.1
  • TAM: install.packages("TAM")  # at least version 4.1
  • eatModel: remotes::install_github("weirichs/eatModel", upgrade="never")
    • Optional software:
  • Positron: free to install from https://positron.posit.co/download.html


    • Schedule

    Recommended readings