MaxDiff allows for a better understanding of the preference order of a set of items and the distance between them. If you are testing between 7 – 200 attributes, aytm's Advanced MaxDiff can help you find the relative preference of those attributes with ease.
If you want to take it a step further and use an individual response value to guide the survey experience, follow along!
Piping Logic: Choice Rank
If you want to ask a follow-up question about a respondent’s top choice in a MaxDiff, you can use piping logic to accomplish this:
[Q#r1] – replace # with the question number of the MaxDiff
“r” refers to response while the number refers to the first, second, third, (etc.) top-ranked choices selected.
For example, we ran a survey on movies in which a MaxDiff asked respondents to rank the feeling they most want to experience when watching a movie. In the follow-up question, we wanted to find out which movie inspires that feeling for them.
In this case, “r1” means their top choice, and “Q12” is the question number of the MaxDiff.
This piping logic will work on any response value for the MaxDiff. If you want to pipe first and last choice or top three choices, you would replace the “1” in the logic statement with the choice position of the response you want to pipe in.
Show/Hide Logic: Position Value
If you want to direct the flow of survey questions based on a respondent's MaxDiff responses, use the logic syntax for position value. For example, if you want to show a question based on a respondent's top-ranked MaxDiff choice, the logic syntax would be:
[show if Q#a#pv=1]
“pv” refers to position value. Don’t forget to replace the # with your question and answer numbers!
Piping with Smart Loops for Optimal Analysis
We recommend utilizing our Smart Loops feature to setup the questions needed to complete the MaxDiff piping. It will provide the most straightforward, cleanest read on your data set – saving you time during your analysis.
In this example MaxDiff, we are asking respondents to evaluate 15 types of candy. In the follow-up question, we want to ask respondents how often they consume their top three choices.
After programming the MaxDiff we will create the Smart Loop. Smart Loops can be selected from the question type menu or by converting an existing question.
Now we will add the attributes--the choices presented to respondents--from our MaxDiff into the Smart Loop table. The list of attributes in the Smart Loop table will act as your “runs,” meaning it will populate a question or set of questions for each run.
Note: the attributes in your Smart Loop must match exactly with the choices in the MaxDiff; any deviation will interfere with the logic.
We’ll change the dropdown at the top to “all/respondent,” so respondents will see all the choices that apply to them based on the “show if” logic we will setup.
The next step is to create the follow-up question or questions within the Smart Loop, using variables.
In this example we want to ask how often they consume their top three choices. The variables in the Smart Loop table are identified by the syntax [candy]. Adding this to your question will pipe each question with the corresponding candy name.
Because we only want to show a respondent this question if they ranked a particular candy first, second, or third in the MaxDiff, we will use [Show if Q1A[run]pv=1 or Q1A[run]pv=2 or Q1A[run]pv=3].
Within a Smart Loop, [run] refers to the attributes populated in the table (which match the MaxDiff choices). In normal terms, this logic will show a respondent a follow-up question if the attribute of the run (e.g., Twix) was selected as a first, second, or third top choice when they responded to the MaxDiff. It will not show a follow-up for any candy types that were not ranked in the top three by the respondent.
By using Smart Loops, we only need to program this once, but the result is that 15 questions will automatically populate to respondents. Now you can preview your survey! Once you are ready you can launch your survey, it's not necessary to convert the SmartLoop unless you need to customize individual questions.
If you need to, you can convert the Smart Loop to 15 individual questions, replacing the [candy] and [run] variables with the syntax referring to your original MaxDiff, as shown below.
In this example Kit Kat is A3 in the MaxDiff and run 3 in the SmartLoop table, [candy] was replaced by Kit Kat and the Show if logic populated the correct answer variant.
Note: if they don’t match, check to make sure your answer list and Smart Loop table list match exactly.
When you use piping logic within a survey, the position value is derived from an on-the-fly estimate of the top choice based on the presented tasks. This means that what might be identified by the platform as the top choice for a respondent on-the-fly might not 100% match their top utility when the final analysis is completed. The calculation is dependent on the variability among the number of times answers were selected. About 90% of the time, on-the-fly estimation will fall within the top 3 utility choices.
Questions? Reach out to us: firstname.lastname@example.org or start a chat!