extract survey number

mez
New Member

Posts: 4

extract survey number Oct 21, 2020 8:37:49 GMT

Quote

Post by mez on Oct 21, 2020 8:37:49 GMT

Hello,

I need 48 participants to do different counterbalanced combinations. I currently have it set up that I would give them a number that they can enter into a survey field which will then direct them to the correctly counterbalanced experiment set up. I am wondering if there is a way to extract what number survey a given survey is? That is can I extract from within the survey that this is the 1st, 2nd or 3rd survey that has been started? My current method is posing problems with the recruitment package we use but if I had the survey number information, it would remove the need for me to email the participant.

Best,
Merryn

PsyToolkit
Administrator

Posts: 545

extract survey number Oct 21, 2020 18:59:24 GMT

Quote

Post by PsyToolkit on Oct 21, 2020 18:59:24 GMT

I am wondering if there is a way to extract what number survey a given survey is? That is can I extract from within the survey that this is the 1st, 2nd or 3rd survey that has been started?

No that is not possible.

I need 48 participants to do different counterbalanced combinations.

For counterbalancing, maybe best to just randomly let people do one the experiments in your survey. If you want a perfect counterbalancing, you can always exclude some participants to make sure you have an equal number in both groups. Of course, you would need to set the criteria for doing so in advance, for example, if you want 24 participants in both conditions, but somehow you end up with 24 in one condition and 27 in the other, you could just leave out the last 3 participants.

Last Edit: Oct 21, 2020 18:59:47 GMT by PsyToolkit

andre
Experienced

Posts: 35

extract survey number Nov 24, 2020 11:03:01 GMT

Quote

Post by andre on Nov 24, 2020 11:03:01 GMT

I was just looking at exactly the same problem, a solution would be great. Access to some form of counter, so that one could say something like

If Participant-Number (a variable provided by PsyToolKit and readable in the Survey) is even, Jump to version A, if it is uneven, Jump to version B.

With smaller participant numbers, random assignment can create quite some uneven distributions (e.g. 6 vs 20 if one is really unlucky). And even if it is not that severe, it seems sub-optimal to just throw away perfectly good datasets.

Even if this "Participant-Number" refers to "started" surveys only (and not necessarily finished/valid datasets), it would ensure a much neater balancing.

PsyToolkit Administrator Posts: 545	extract survey number Dec 1, 2020 23:27:10 GMT Quote Select Post Deselect Post Link to Post Member Give Gift Back to Top Post by PsyToolkit on Dec 1, 2020 23:27:10 GMT I agree, I will put it on the wish list.

extract survey number

Post by mez on Oct 21, 2020 8:37:49 GMT

Post by PsyToolkit on Oct 21, 2020 18:59:24 GMT

Post by andre on Nov 24, 2020 11:03:01 GMT

Post by PsyToolkit on Dec 1, 2020 23:27:10 GMT