How to Use Qualtrics MaxDiff Variables in Q

Q can analyze MaxDiff variables from a variety of survey platforms, but depending on how they have been set up, you may need to reformat them first. One such example is the output from Qualtrics' dedicated MaxDiff module.

In Q, MaxDiff analysis requires an experimental design, and separate Best and Worst variables for each question or task where the alternative labels are the category options.

Typically this is programmed in Qualtrics as a series of matrix questions (which may include piped alternatives), whereby the variables are labelled according to which alternatives were shown for each question or task, and the category labels represent the Best or Worst selections. 

This means the variables need to be reformatted and the design created in order to work with our built-in analysis tools.

Requirements

• A data set that has been set up using Qualtrics' dedicated MaxDiff module, or where the alternatives are the variable labels and the Best and Worst selections are the category labels, as per the above example.
• The QScript for converting Qualtrics MaxDiff data saved here.

Method

Part 1 – Prepare your data

1. Go to your Variables and Questions tab.
2. Make sure your MaxDiff variables have been set up as a set of Pick One - Multi questions, one for each task or iteration in the design. If your design asks each respondent to evaluate 4 sets of alternatives 6 times, then there should be 6 sets of Pick One - Multi variables made up of 4 variables each in your project. If they are not already grouped, you will need to make sure the Variable Type is set to Categorical, then select all the variables for each iteration, right-click > Set Question > OK.
3. If the alternatives shown to each respondent are stored in a different set of questions:
   1. then you also need to ensure these have likewise been grouped in the same way. When designed using the Qualtrics MaxDiff module, these will be named in the format of A1.1_MAXDIFF.
      They should look like this:
   2. You should have a design version variable for the maxdiff (seen above as vers_MAXDIFF). This should have a Question Type of Pick One or Number.

Part 2 – Run the QScript

1. Download the QScript zip file linked in the above Requirements section and extract the .QScript file.
2. In Q, go to Automate > Run QScript (Macro) From File.
3. Select your downloaded .QScript file and press Open.
   
4. Select the set of questions that contain the MaxDiff selections then press OK.
5. Next, you will be prompted to confirm whether the alternatives that were shown to respondents are stored in a different set of variables. If you select No, the script will run based on a non-randomized single version design. If you select Yes, the script will run based on a randomized multiple version design.
6. [OPTIONAL] If Yes, you will be prompted to select your design version variable.
7. [OPTIONAL] If Yes, you be prompted to select the MaxDiff questions that store what alternatives were show to the respondents.
8. The script will generate a folder called MaxDiff Design which includes your design output.
9. The script will also generate a folder called MaxDiff Variables which displays the newly created variables as a table.

Your design should look like this:

And the table of your transformed variables should look like this:

Your MaxDiff data is now ready to be analyzed in Q, choose a method below.

Next

How to Do MaxDiff Latent Class Analysis in Q

How to Use Hierarchical Bayes for MaxDiff in Q