Rank order question - how to export data by ranking (not by things being ranked)?

@JB_TDSB ,

You’ve nailed the problem I’ve always had with Rank order questions on the head. That is why I personally avoid using this question type. Rather, I prefer to use Multiple Choice Questions: Instead of using the Rank Order question type, you can create a series of multiple choice questions with a single choice allowed. Each question can represent one ranking position, and the options can be the items to rank. For example:

With this setup, you will have individual responses for each ranking position, and you can export the data accordingly.

• Question 1: "Rank 1" (Options: Option 1, Option 2, Option 3, Option 4, Option 5)
• Question 2: "Rank 2" (Options: Option 1, Option 2, Option 3, Option 4, Option 5)
• Question 3: "Rank 3" (Options: Option 1, Option 2, Option 3, Option 4, Option 5)
• ...

This workaround provides a way to obtain the ranking data directly.

There is a way you can also create custom embedded data fields to store the rankings selected by participants, but it requires Javascript. Let me know if you are interested in that approach.

Thank you for the reply.  I really appreciate it. 

If I follow the multiple choice method you have outlined, I gather there is no preventative mechanism to stop someone from listing 3 different options as their first choice.  I assume the work around to this would be to have a formula highlight repeats within in individual votes.

Is there a tool I’ve overlooked that would alert or prevent a user that they input 3 first choice options, which isn’t permitted, prior to them submitting? 

Javascript is beyond my knowledge, but thank you for the willingness to share.  

Hi, one other option to consider that would give you the data in the format you're looking for with a similar respondent experience is to use a transposed matrix table. The only downside is that columns/scale points are not mutually exclusive like they are with the radio rank order, though Custom Validation can be added so that the Count of each column must be equal to 1 in order to proceed.

To give it a try, create a Matrix table and set the statements to be "1st choice", "2nd choice", and "3rd choice". Then set the scale points to be "Red", "Blue", "Yellow". Next, enable the "Transpose table" setting so that the statements appear as columns to the respondent. 

Finally, click "Add validation" under response requirements and set it up like the below:

@JB_TDSB , 

Yes, you can use the “Carry Answers Forward” functionality and bring forward the “Unanswered options” like this: 

That way you can ask those questions and skinny it down to 1st, 2nd, 3rd, etc. 

If you stick with the original rank order question, could you use survey flow to set the embedded variables for Q1_1st, Q1_2nd, Q1_3rd etc to be the values you want?

@MikeW nothing that is very easy although there are doable (unpleasant options). Qualtrics will only pull in the attribute selected for First and Last choice using piped text:

All other piped text options will either show you the attribute label without any logic of why it’s being inserted or the number chosen that goes with the label. To flip it so the variable is First, Second, Third… and the value is the attribute you would need to create branching logic for each variable and each possible rank. So in my example I had three choices and setting this up for the first attribute only to save into First, Second, or Third it would look like:

This would need repeated for every variable as many times as variables you have (e.g. with 5 choices you will have 25 branches).

bstrahin wrote:

@MikeW nothing that is very easy although there are doable (unpleasant options). Qualtrics will only pull in the attribute selected for First and Last choice using piped text:

All other piped text options will either show you the attribute label without any logic of why it’s being inserted or the number chosen that goes with the label. To flip it so the variable is First, Second, Third… and the value is the attribute you would need to create branching logic for each variable and each possible rank. So in my example I had three choices and setting this up for the first attribute only to save into First, Second, or Third it would look like:

This would need repeated for every variable as many times as variables you have (e.g. with 5 choices you will have 25 branches).

sorry - that does sound pretty horrible