How would you setup a study like this one? | XM Community
Skip to main content
I am setting up a study that has four videos, each followed by a task. The hypothesis is related with the effect of each video on the performance of each consequent task.



The tasks are very similar but not identical to avoid a learning effect and I would like to randomize the way both the videos and the tasks appear. If the same task always follows the same video, the differences in performance could be attributed to difference in the tasks themselves (we piloted the difficulty of the tasks to check that they were similar, but still…)



The aim would be that each person sees one of the four videos, then perform one of the four tasks, continue with a second video and a second task, until completing the four videos and the four tasks.



The challenge is how to setup a flow for a design like this one, ensuring that the order of the videos and the order of the tasks are randomized without repeating.



Any idea on how to do it?
Use a survey flow randomizer to randomize your tasks into embedded variables. To avoid confusion, lets call the tasks A, B, C, and D. You have 24 possible orders, so you would create 24 embedded data blocks and have the randomizer pick 1 (you could do fewer blocks if you don't care about hitting all the combinations). The first embedded data block would look like the following, then the others would have the numbers in different orders:

```

taskA = 1

taskB = 2

taskC = 3

taskD = 4

```

Then setup a loop and merge block with your four videos urls as Field 1, and randomize the loop. The first question in your loop is descriptive text and it displays the piped in video in Field 1. The following questions are your tasks. For each task question, add display logic like this:

```

If taskA Is Equal to ${lm://CurrentLoopNumber}

```
That answer is perfect! Thank you very much for your help.



I can set up the study following your suggestions

Leave a Reply


Terms & Conditions