You need to fit an IRT model for a course or a thesis, check that it holds up, and explain it in figures and plain words. mirtverse does the bookkeeping so you can spend your time on the interpretation.
Your path in three steps
1. Learn the ideas with the book and the labs. The free online book Item Response Theory with mirt and mirtverse has one chapter per week of a typical course, from classical item analysis to reporting results. The interactive labs let you practise on realistic data:
2. Run your own analysis. Choose a model, read what it means, and check the items:
sel <- select_model(mydata) # compares Rasch, 2PL and 3PL and explains the choice
mod <- sel$best
interpret(mod) # what the results mean, in plain language
item_table(mod) # item statistics, fit and flags in one tablePrefer clicking? launch_workbench() does the same steps
with buttons and writes the R code for you, which you can paste into
your thesis appendix.
3. Make figures and a report. Every plot is a ggplot2 object, so you can restyle it for your university’s template:
gg_wright(mod) # person-item map
gg_info(mod) # test information
mv_report(sel, file = "thesis-analysis.html") # everything in one fileThe Gallery shows what each figure looks like.
Keep going
- Glossary: Nearly 90 IRT terms in plain language.
- Cheat sheet: The main functions on two printable pages.
- Which IRT model?: A one-page flowchart from your item scores to a model and the code that fits it.
-
Coming from ltm, eRm or TAM?
mirtify()rewrites your old code as mirt code.
Writing it up
The thesis kit walks from data to a
finished Methods paragraph. cite_methods(mod) drafts that
paragraph in APA style, with the software versions, model, estimation,
fit, reliability and references, so you only have to check and edit
it:
cite_methods(sel, sample = "undergraduate students", file = "methods.docx")Cite mirt (Chalmers, 2012) and mirtverse in your methods section;
cite_methods() adds both references for you.