All Research Guides

A Guide From Domandata

How to Choose Conjoint Analysis Software

Written by Domandata, which makes a survey platform. We name our product where it fits and say where another tool fits better. Claims about other tools link to their own documentation, checked ; confirm anything decisive with the vendor or your institution.

Conjoint software splits into two traditions that answer different questions. Market-research suites estimate each respondent's utilities with hierarchical Bayes and simulate market shares. Social scientists, following Hainmueller, Hopkins, and Yamamoto, usually randomize attributes fully and estimate average marginal component effects (AMCEs) or marginal means.

Choosing the wrong tradition gives you outputs that do not match your estimand, so this guide starts there, then compares the tools in each.

The short answer

For AMCEs and marginal means, design a fully randomized conjoint in a survey tool that records every displayed profile, then analyze in R with cjoint or similar code. Domandata builds this kind of conjoint natively; the free Conjoint Survey Design Tool does it inside Qualtrics with generated JavaScript. For hierarchical Bayes utilities and market simulation, use Sawtooth Software, Conjointly, or Qualtrics Conjoint, which is a separate purchase.

What to decide before picking conjoint software

  1. 1

    Your estimand

    AMCEs and marginal means describe average effects of attribute levels across respondents and need only randomization and a regression. Individual utilities and preference shares need hierarchical Bayes models and suit prediction and market simulation.

  2. 2

    Randomization and restrictions

    Full randomization keeps the AMCE straightforward to estimate. If some combinations are impossible, you need prohibited combinations, and the analysis must account for them.

  3. 3

    Recording what was shown

    Your data must include every attribute level of every profile in every task, plus the alternative position and the choice. Without it you cannot estimate anything, so check the export in a pilot.

  4. 4

    Task design

    Number of tasks and alternatives, an opt-out option, holdout tasks, and attention checks all affect data quality. A forced choice with no opt-out can bias results when respondents dislike every profile.

  5. 5

    Power

    Rare levels, many attributes, and subgroup comparisons all demand larger samples. Run a power calculation for your smallest effect of interest before recruiting.

  6. 6

    The analysis path

    Know which software estimates your model. Commercial suites build analysis in; academic workflows usually export long-format data to R.

Conjoint tools by approach

AMCEs and marginal means for social science

Domandata builds randomized conjoints with prohibited combinations, opt-outs, holdouts, image levels, cross-respondent balancing, and AMCE power planning, and exports every displayed profile. The Conjoint Survey Design Tool generates a randomizer for Qualtrics. cjoint estimates AMCEs in R from either.

DomandataOur product

Browser-based survey platform built for academic social-science research: survey experiments, conjoint, and analysis-ready data.

Pricing model
Free plan for drafting and previewing; fielding a study, Flow logic, and exports need a paid individual or team plan with published self-serve pricing.
Where it runs
Cloud only, on U.S.-based infrastructure providers.

Strengths

  • Randomized treatment arms (weighted, stratified, or balanced) with the assigned condition saved in every export.
  • Conjoint, MaxDiff, list experiments, and reaction-time tasks built in, without custom JavaScript.
  • Native SPSS (.sav) and Stata (.dta) exports with a generated codebook, plus Qualtrics .qsf import.
  • Links OSF preregistrations and deposits replication materials to Harvard Dataverse or Zenodo.

Watch out for

  • Does not sign HIPAA Business Associate Agreements, so it is not suitable for studies that require one.
  • Cannot be self-hosted, and the data-hosting region is not fixed by contract.
  • The free plan cannot collect responses; fielding a study requires a paid plan.
  • A newer, smaller vendor: its SOC 2 Type I audit is still in progress, which some IT reviews ask for.

Sources:Domandata PricingDomandata Trust Center

Conjoint Survey Design Tool

Free program from the authors of the standard AMCE approach that designs fully randomized conjoints and generates a randomizer for Qualtrics.

Pricing model
Free and open source (GPL-3.0).
Where it runs
Runs from Python source on your computer; output is pasted into Qualtrics.

Strengths

  • A standard tool in political science, actively maintained.
  • Pairs directly with the cjoint package for analysis.

Watch out for

  • Requires running Python and pasting JavaScript into Qualtrics.
  • Using it with another survey platform means adapting the generated code yourself.

Sources:Conjoint Survey Design Tool on GitHub

cjoint (R package)

R implementation of the Average Marginal Component Effect (AMCE) estimator for conjoint experiments.

Pricing model
Free on CRAN.
Where it runs
Runs in R on your computer.

Strengths

  • Implements the AMCE estimator from Hainmueller, Hopkins, and Yamamoto.
  • Maintained on CRAN, with a 2026 release.

Watch out for

  • Analysis only; you design and field the conjoint elsewhere.

Sources:cjoint on CRAN

Hierarchical Bayes utilities and market simulation

These suites estimate individual utilities, importance, and preference shares with built-in analysis. Sawtooth is the most established, Conjointly is web-based with an academic waiver, and Qualtrics Conjoint adds conjoint to an existing Qualtrics account.

Sawtooth Software

Long-established commercial conjoint suite covering choice-based conjoint, adaptive conjoint, and MaxDiff, with hierarchical Bayes estimation.

Pricing model
Annual per-user subscriptions, with discounted academic research licenses, free teaching licenses, and free grant licenses for graduate students.
Where it runs
Web-based Discover or the Windows desktop Lighthouse Studio, with free hosting on Sawtooth servers or self-hosting.

Strengths

  • Mature hierarchical Bayes estimation and market simulation.
  • Advanced designs such as adaptive choice-based conjoint.
  • Free licenses for many students and instructors.

Watch out for

  • Lighthouse Studio runs only on Windows.
  • Oriented to utilities and market simulation rather than the AMCE estimands common in political science.

Sources:Sawtooth Academic PricingSawtooth Academic Grants

Conjointly

Web platform for conjoint, MaxDiff, and pricing studies with built-in analysis.

Pricing model
Free basic tier with tight limits on advanced methods, annual paid licenses, and a free one-year academic fee waiver by application.
Where it runs
Cloud.

Strengths

  • Analysis is built in, so results are available without coding.
  • You can bring your own respondents at no cost.

Watch out for

  • Built for market research; the academic waiver excludes sampling and generally includes no support.

Sources:Conjointly PricingConjointly for Academics

Qualtrics Conjoint

Qualtrics' conjoint project type, which estimates individual utilities with hierarchical Bayes and reports importance and preference shares.

Pricing model
An additional purchase on top of a Qualtrics license, arranged through an account executive.
Where it runs
Cloud, inside Qualtrics.

Strengths

  • Built-in design, fielding, and reporting in one place.
  • Exports individual utilities, importance, and preference shares.

Watch out for

  • Not included in a standard license.
  • Documented outputs are hierarchical Bayes utilities; AMCEs and marginal means need separate analysis.

Sources:Analyze Conjoint DataExport Conjoint Data

Custom or interactive designs

If your conjoint is embedded in an interactive experiment or needs fully custom logic, oTree's examples show how to build one in Python.

oTree

Python framework for interactive, multiplayer experiments such as economic games and markets, which also supports surveys.

Pricing model
Free under an MIT-style license that asks you to cite the oTree paper in publications.
Where it runs
Self-hosted on Heroku or your own server.

Strengths

  • Real-time interaction between participants, which survey platforms generally do not offer.
  • Free and open source.

Watch out for

  • Requires Python programming.
  • You set up and run the server.

Sources:oTreeoTree Server Setup

Pilot checklist for a conjoint

  • Every attribute level appears in pilot data at roughly the expected rate
  • Prohibited combinations never appear
  • Each row records task, alternative position, all levels, and the choice
  • Attribute order behaves as designed, fixed or randomized per respondent
  • The opt-out, holdout, and attention tasks are coded distinctly
  • Your analysis script runs on the pilot export without reshaping by hand

Frequently Asked Questions

What is the best free conjoint analysis software for academics?

For AMCE-style studies, the Conjoint Survey Design Tool and the cjoint R package are free, though the tool targets Qualtrics. Conjointly offers a free one-year academic waiver by application, and Sawtooth Software offers free grant licenses to many graduate students and instructors.

Does Qualtrics Conjoint give AMCEs?

Qualtrics documents its conjoint outputs as hierarchical Bayes utilities, attribute importance, and preference shares. Researchers who need AMCEs or marginal means typically export the displayed profiles and choices and estimate them in R.

Is the cregg package still available?

The cregg package, which estimates AMCEs and marginal means, was archived from CRAN in September 2024 and is no longer updated there. It can still be installed from its GitHub repository; cjoint remains on CRAN with a 2026 release.

How many attributes and tasks should a conjoint have?

Include only the attributes your hypotheses need, since each extra attribute adds burden and dilutes exposure to every level. Set the number of tasks from a pilot of respondent burden and from the power calculation for your rarest level, rather than copying another study's design.

Keep Reading

Conjoint Analysis Survey Design

Attributes, randomization, task structure, and exports in depth.

Build a Conjoint in Domandata

Configure attributes, levels, tasks, balance, and holdouts.

What Social-Science Surveys Need

Requirements for social-science designs, with pilot tests.

Choosing a Recruitment Platform

Where to find respondents for your conjoint.