Preface

Book cover: an engraved owl on a branch above the title Item Response Theory with mirt and mirtverse

This book is a set of extended course notes on item response theory (IRT). It is written for instructors who teach measurement with R and for the students in their courses. Every chapter mixes a short explanation of the ideas with worked examples that you can run line by line, using the mirt package (Chalmers 2012) for estimation and mirtverse for model choice, tidy tables, graphics and plain-language interpretation.

How the book is organised

The twelve chapters follow the twelve weeks of the mirtverse course map, so an instructor can assign one chapter a week and pair it with the matching lab.

Table 0.1: The twelve chapters and their matching labs.
Chapter Topic Matching lab
1 Measurement, scoring and classical item analysis Lab 1 (first two sections)
2 The Rasch, 2PL and 3PL models
3 Estimation and comparing models
4 Dimensionality and local independence Lab 1
5 Item fit, person fit and item quality
6 Information, precision and reliability
7 Scoring people Lab 2
8 Polytomous items and open-ended responses Lab 6
9 Differential item functioning and fairness Lab 3
10 Linking and equating Lab 4
11 Item banks, test assembly and adaptive testing Lab 5
12 Cut scores, reporting and communicating results Final project

A second part, AI in measurement, is for readers who know IRT and want to learn how machine learning and large language models fit into psychometrics. It goes with the mirtverse AI Hub and grows as the hub does.

Table 0.2: The chapters on AI in measurement.
Chapter Topic Matching lab
13 Machine learning through an IRT lens
14 What item text tells us Lab 8
15 Generating items Lab 9
16 Automated scoring Lab 7
17 Large language models as a colleague
18 Validity, fairness and governance of AI Lab 10

A third part, Knowledge graphs, describes what a test measures as a graph of concepts, prerequisites and items, and turns the graph into IRT models. Chapter 19 builds and checks the graphs, and Chapter 20 follows students’ mastery of the concepts over time with knowledge tracing.

Every chapter has the same parts:

  • Learning goals: What you should be able to do after the chapter.
  • Key terms: Links to the mirtverse glossary.
  • The ideas: A short explanation, with formulas only where they help.
  • Example 1 and Example 2: Two worked analyses on different data, so you see the same method in two settings.
  • Your turn: Exercises with solutions. In the online edition the solutions are folded away; click to open them.
  • In the lab: The interactive mirtverse lab that goes with the chapter.
  • Further reading.

The examples deliberately use different data from the labs. The labs hide problems in their data for students to find, and the book does not give those away.

Before you start

You need R (version 4.1 or later), mirt (version 1.48 or later) and mirtverse:

install.packages(c("mirt", "learnr"))
install.packages("mirtverse", repos = c("https://okanbulut.r-universe.dev",
                                        "https://cloud.r-project.org"))
library(mirtverse)

Loading mirtverse also loads mirt, so mirt(), fscores(), itemfit() and the mirt datasets are available straight away. Every example runs in a few seconds on a laptop, and no internet connection is needed once the packages are installed.

Students who are new to R can follow much of the book without writing code. The workbench (launch_workbench()) runs the same analyses by pointing and clicking and shows the R code for every step, and mv_ask() writes a script from a plain-language request.

Other teaching materials

  • Teaching labs: Ten interactive learnr labs with an instructor guide and answer key.
  • Course map: A 12-week syllabus with readings, functions, labs and assessment ideas.
  • Glossary of IRT terms.
  • Cheat sheet of the main functions, on two printable pages.
  • Assessment pipeline: A test’s full life cycle from blueprint to bank monitoring.

Using and adapting the book

The source of every chapter is an R Markdown file in the book/ folder of the mirtverse repository. Each example loads its data in one clearly labelled chunk, so you can swap in your own data and rebuild the chapter. Appendix B explains how.

Other formats

The book is free to read online at https://www.okanbulut.com/mirtverse/book/. It is also available as a PDF for printing and an EPUB for e-readers; in the online edition, use the download button in the toolbar. All three are built from the same source each time the book changes.

How to cite this book

If you use the book in a course or in your work, please cite it as:

Bulut, O. (2026). Item response theory with mirt and mirtverse: Notes for teaching and learning. Zenodo. https://doi.org/10.5281/zenodo.23885605

DOI

This is version 1.1 of the book. The DOI above always points to the latest edition. Each archived edition also has its own DOI, listed on the Zenodo record; cite that one if you need to refer to the exact text you read (version 1.0 is 10.5281/zenodo.23885606).

@book{bulut2026irt,
  author    = {Okan Bulut},
  title     = {Item Response Theory with mirt and mirtverse: Notes for Teaching and Learning},
  year      = {2026},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.23885605},
  url       = {https://www.okanbulut.com/mirtverse/book/}
}

Please also cite the software you use: mirt (Chalmers 2012) and mirtverse. Running citation("mirtverse") in R gives the mirtverse reference with its DOI.

Licence

© 2026 Okan Bulut. The text and figures of this book are licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International licence (CC BY-NC-SA 4.0): You may share and adapt them for non-commercial purposes, such as teaching, as long as you give credit and share your adaptations under the same licence. The R code in the book is licensed under GPL-3, like mirtverse, so you can reuse it freely in your own analyses.

Acknowledgements

This book stands on the work of many people. Phil Chalmers wrote and maintains mirt, which does all of the estimation in these pages. The book was written with R Markdown and bookdown, by Yihui Xie and colleagues, and the figures were drawn with ggplot2. My thanks to all of them, and to the R community that makes open teaching materials like this possible.

References

Chalmers, R. Philip. 2012. “mirt: A Multidimensional Item Response Theory Package for the R Environment.” Journal of Statistical Software 48 (6): 1–29. https://doi.org/10.18637/jss.v048.i06.