|
|
Post by joebutler on May 6, 2020 19:46:46 GMT
Hello,
I'm interested in running an online study which gives a couple of questionnaires and then administers two tasks (a digit span task and the attention network task - both I assume will be easy to implement when I've learned how the syntax works).
The main part of the study will actually be the questionnaires and then the behavioural data will be a nice-to-have, so if subjects quit part way through the tasks that will not be too much of a problem (or should I say, I'm resigned to a high drop out rate!).
The only questions I have is, is the data still be available if a subject prematurely finishes the task (as I expect high drop out rate during the cognitive tasks)? And also is it possible to administer two cognitive experiments, or would I work around this within one experiment? (e.g. code the different tasks as different blocks).
And apologies if this seems like an odd question, but I saw that on another package for running online experiments that if the study is not completed, then the data is lost.
Thanks for your time,
Joe
|
|
|
|
Post by PsyToolkit on May 7, 2020 8:44:57 GMT
Incompleted data are still stored and available, but I generally would not recommend to use incomplete data for serveral reasons:
1) Participants might stop because they realize they actually do not want to answer the questions. Make sure that if you analyze incomplete surveys, that you inform your participants about this in the welcome screen (in case IRB/ethics approval apply).
2) Participants might not have taken the survey seriously and just wanted to try it out -- you will not know for sure, again, the question is whether the incomplete surveys contain valid data.
Also, I would try to make sure people can finish the study. Why not just give people and option, something like this: "Now you have the option to do a test which measures your error rates, it takes 3 minutes, do you want to participate this or skip?".
Yes, you can embed as many experiments as you want.
Generally, with online studies, people drop out for various reasons (I based this on my own experience):
- It is too boring (giving responses without getting feedback)
- It is too long (anything longer than 15 minutes is quite long)
- It is too difficult (this applies to experiments or poorly written questionnaires)
- It is too repetitive (the same questions seem to be asked over and over again)
- Lack of motivation or incentive (give some feedback or motivating message, that can be done)
|
|
|
|
Post by joebutler on May 7, 2020 16:49:03 GMT
Thank you very much for your considerate answer. I've mentioned these points to my collaborator and we've thought about how to handle the task ethically. Perhaps by asking participants to generate unique identifier at the start of the task, and then emailing us with this, if they wish to withdraw (so the assumption is they are fine with us keeping partial data).
The partial data to us would be the full questionnaire data (we hope everyone will complete this). We do expect high drop out with the Attention Network Task, as it takes around twenty minutes (it's basically a flanker task, with the flanker stimuli preceded by spatial/alerting cues - although I'm sure you are already familiar with the task). There's a few things we don't want to mention on a public forum, but we hope that people are basically motivated to contribute to the science because of the topic.
I'm going to spend this weekend getting familiar with the syntax, but I must say it seems very good so far.
Thanks again,
Joe
|
|
|
|
Post by PsyToolkit on May 7, 2020 22:15:29 GMT
Hi Joe, you can also create a unique code automatically at the beginning of your study. There is a way to do this at the end, but it seems in your case you want to give something like that. You can do this as follows:
l: my_random_code t: set - random 10000 99999
l: give_code t: info q: We give you a special number now, which is: {$my_random_code} Write this down. If you want to withdraw from this survey for any reason, please send this number to some_email@email.com. Even if you do not complete the whole survey, we will use those data you have submitted so far (unless you write us with the number telling us that you do not want that)
l: etcetctc t: radio q: What do you like to snack
- Chips - Popcorn - Nuts
|
|
