The Conjoint Survey Design Tool assists researchers in creating multi-dimensional choice experiments that can be readily incorporated into any pre-existing web survey software (such as Qualtrics). Conjoint analysis is a type of survey experiment often used by market researchers to measure consumer preferences over a variety of product attributes. Hainmueller, Hopkins and Yamamoto (2014) demonstrate the value of this design for political science applications. Conjoint experiments present respondents with a choice among set of profiles composed of multiple randomly assigned attributes. This approach allows researchers to estimate the effect of each individual component on the probability that the respondent will choose a profile. This software tool is designed as a companion to Hainmueller, Hopkins and Yamamoto (2014), providing a graphical user interface for generating conjoint experiments.
Hainmueller, Jens., Hopkins, Daniel J., Yamamoto, Teppei. (2014). Causal Inference in Conjoint Analysis: Understanding Multidimensional Choices Via Stated Preference Experiments. Political Analysis 22(1), 1-30
There are two ways to run the Conjoint Survey Design Tool (SDT).
For Windows users, an executable binary is available for download. This is a self-contained package that requires no additional libraries to function. To use the SDT, download the conjointSDT.exe
file to the desired location and run the executable.
A sample survey file, immigrant_demo.sdt
, is included in the Demos folder.
Mac OSX and Linux users do not have standalone executables and need to use the Python source files. Windows users with a Python installation can also use the source package. This is advisable as the binary package is much larger than the source since it includes the necessary Python libraries and interpreter.
The Conjoint SDT is written for Python 3 and requires Python 3.6 or greater. To obtain the most recent release of Python 3, visit https://www.python.org/downloads/
Mac OSX users should make sure that they have the most recent version of the TCL/TK library installed. If you installed Python. Consult https://www.python.org/download/mac/tcltk/ for more information.
To run the Conjoint SDT from Python source, download the conjointSDT.py
to the desired directory and run the file through the Python interpreter (this can be done through the command line by calling python conjointSDT.py
or python3 conjointSDT.py
if your installation distinguishes between versions 2 and 3 of python).
A sample survey file, immigrant_demo.sdt
, is included in the Demos folder.
Please consult the conjoint_sdt_manual.pdf
file located in the Manual folder for detailed instructions on how to use the GUI design tool.
After implementing the experiment and collecting the result, researchers can use the cjoint
R package for analysis and visualization of the results. This package can be installed directly from CRAN. The most current documentation can be found at https://cran.r-project.org/web/packages/cjoint/index.html
.
Version 3.0 - January 10, 2022
Added a JavaScript randomizer to supplement the existing PHP randomizer. Users can now export JavaScript snippet that can be placed directly within a Qualtrics question to generate the conjoint tasks for a respondent. This no longer requires users to have a second server to host the PHP script and allows the randomization be carried out entirely within Qualtrics.
Added a feature to allow users to generate designs that prevent identical profiles from appearing in a single task (off by default).
Version 2.0 - July 24, 2019
Ported over to Python 3. Fixed various compatability issues that had accumulated over the last several years.
Version 1.3 (BETA) - May 16, 2014
Fixed significant bug in Qualtrics question templates. Qualtrics template files will not longer include choice radio buttons. Do NOT use radio buttons within a Descriptive Text item to obtain responses as Qualtrics will not record any data passed through a Descriptive Text item. Instead, use a Multiple Choice item and create choices that correspond to each profile.
Version 1.2 (BETA) - May 5, 2014
Added "Export to R" feature to allow designs to be exported to the (forthcoming) conjoint R package for estimating component effects
Version 1.1 (BETA) - November 16, 2013
Added automatic question templates for use in Qualtrics
Version 1.0 (BETA) - September 23, 2013
First release
If you have further questions about using the Conjoint SDT or wish to report a bug, please do not hesitate to contact Anton Strezhnev at astrezhnev@uchicago.edu.
GNU General Public License 3.0
Special thanks to Katarina Jensen for assistance in porting the old Python 2 code to be compatible with Python 3. Thanks to everyone who has contributed with bug reports and feature suggestions.