MaxDiff

Open tools for creating and analyzing MaxDiff studies.

Recommended default: Use count-based analysis with bootstrap confidence intervals for most MaxDiff studies. Need Hierarchical Bayes (HB) or individual-level utilities? Use the Windows desktop version from GitHub Releases.

MaxDiff Analysis Tool

Run the browser-based analysis tool. Your data stays local in your browser. Supports count-based scores and bootstrap confidence intervals.

Open analysis tool

MaxDiff Design Tool

Create balanced MaxDiff designs in the browser.

Open design tool

Windows Desktop Tool

Download the desktop executable for a local Windows app. It defaults to count-based analysis and can enable HB when individual-level utilities are needed.

Get desktop version

Python Package

Use the analysis library directly in scripts and notebooks: pip install maxdiff. Count-based scoring, bootstrap CIs, and optional Hierarchical Bayes via pip install maxdiff[hb].

View on PyPI

Further Reading

“MaxDiff's max myths” (Carl Jago, Quirk's): on when MaxDiff is, and isn't, the right tool for the job. Worth reading before reaching for any of the tools above.

Read the article

Source code and downloads: github.com/milkywaycj/maxdiff