Batch session update after each participant

trig

Hi,

I am wondering if (and if yes, how) it is possible to update the batch session data after every participant and I guess my question is twofold.

Suppose I have an experiment with two between-subjects conditions, in each of which I want to have 5 participants in the end and I am doing condition assignment using Batch Session Data (i.e., the batch session data would look something like

{

 "conditions": [  "A",  "A",  "A",  "A",  "A",  "B",  "B",  "B",  "B", "B" ]

}

1) Places in the experiment are obviously taken away also from participants who start but then abort the study (or just never finish it). Is there a way to put that condition back into the batch session data, for instance if the experiment is not finished within a given timeframe?

2) Of course, I would only want 5 "good" participants per condition, so only those who passed an attention check and with full datasets, for instance. I could implement my attention check script within my experiment, and again: if a participant had failed the attention check, could I replace that condition again in the Batch Session Data?

My overall goal is to have equal-sized conditions more easily, rather than frantically checking the data of all participants so far towards the end of a data collection and manually changing the available spaces in the experiment.

Thanks!

elisa

Hi Trig,

The quick answer is that it’s all possible, but I wouldn’t (necessarily) recommend it.

Within your script you can implement whatever you want. To modify the batch session data you’re most likely going to need jatos.batchSession.add. There’s a pretty detailed description of how to use this function and other related ones in the docs, but do ask again if you have more specific questions once you start writing this.

http://www.jatos.org/jatos.js-Reference.html#jatosbatchsessionadd

However, in my experience, “frantically analyzing the results” is not too bad. Even if (say) you’ve preregistered your exclusion criteria, It’s still pretty difficult to know in advance al possible ways in which an experiment goes wrong (especially online) and sometimes implementing this to work automatically takes more time than what it saves. You could always aim to collect a bit more than the 5 participants that you strictly need. And in any case you will have to look at your data. Granted, maybe not during data collection, but again in my experience that’s doable.

Anyway, just an opinion. Feel free to ignore it and ask again if you have trouble or more questions implementing it.

best

Elisa

trig

Thanks Elisa!

I will look more closely at the function description. And: I might have overstated my goal--I don't just want to do all proper checks during the data collection but check the data every now and then anyway. My main concern was that--especially when collecting data via Prolific--things can get a bit complicated towards the end of a data collection, and I think the batchSession.add might at least reduce some of the troubles :)