By cfinlm - 4/7/2015
Hello,
I'm using Inquisit Web 4 with 4 sequential scripts (1. info questions, 2. IAT, 3. demo questions, 4. likert questions) to administer a study to participants. So far about 50 participants have completed the study and I've noticed 2 of the participants don't have data for all of the scripts. It's not just that they aren't participating, it appears from the web log that they were never administered these scripts in the first place. One participant did not receive the IAT script even though I asked her to try it multiple times. Another participant (who was having trouble accessing the study at all with Chrome41 but was able to when she switched to IE11) wasn't administered the demographic questions script. Since it's happening to such a small portion of participants it appears to be almost a random anomaly. Any ideas why this is happening?
Thanks!
Carli
|
By Dave - 4/7/2015
It's impossible to say without further details. It might be some random glitch, although this would be the first time I've heard about it. The respective participants might have (mistakenly) quit the respective script(s) by pressing CTRL+Q.
Or you might have some error in your scripts, perhaps involving inappropriate /stop or /skip conditions, or inappropriate condition assignment (see below).
> it appears from the web log that they were never administered these scripts in the first place
Depending on how those scripts are set up / how you assign conditions (via the /subjects and /groupassignment attributes), there may be certain combinations where those conditions simply *do not apply* and hence the script isn't administered. As a simple example, suppose you have two <expt> elements set up like this:
<expt>
/subjects = (1 of 2)
/groupassignment = subjectnumber
[...]
</expt>
<expt>
/subjects = (2 of 2)
/groupassignment = groupnumber
[...]
</expt>
The first <expt> assigns based on *subjectnumber*, the second one uses *groupnumber*. The consequence is this: If an *odd* numerical subject id is provided, one should receive the blocks defined in the first <expt>. If the random groupnumber is *even*, the second <expt> might be run in addition. Now suppose the subject number is *even*. Consequently, the first <expt> doesn't apply then. Further suppose the groupnumber is *odd*, i.e., the second <expt> doesn't apply either. As a matter of
fact, in this case *no condition* applies, and the script would not be run at all.
|
By cfinlm - 4/7/2015
Hi Dave,
Thanks for getting back. The participants didn't push control q on these items. I've checked all my scripts and there is only one /skip condition (in the IAT):
<block summary>
/skip = [values.showsummaryfeedback == false]
/ trials = [1=summary]
/ recorddata = false
</block>
Group assignment shouldn't be the problem either:
<expt>
/subjects = (1 of 2)
/groupassignment = groupnumber
/ blocks = [1=targetcompatiblepractice; 2=attributepractice; 3=compatibletest1; 4=compatibletestinstructions;
5=compatibletest2; 6=targetincompatiblepractice; 7=incompatibletest1; 8=incompatibletestinstructions;
9=incompatibletest2; 10=summary]
/onexptend = [values.completed = 1]
</expt>
<expt>
/subjects = (2 of 2)
/groupassignment = groupnumber
/ blocks = [1=targetincompatiblepractice; 2=attributepractice; 3=incompatibletest1; 4=incompatibletestinstructions;
5=incompatibletest2; 6=targetcompatiblepractice; 7=compatibletest1; 8=compatibletestinstructions;
9=compatibletest2; 10=summary]
/onexptend = [values.completed = 1]
</expt>
Even if the skip or group assignment items were the issue it doesn't make sense to me why one would skip the IAT script (in which both of these are located) and one would skip the demographic question script.
Might this issue be related to some factor on their computers? Any other ideas? I'd be happy to send you my scripts if you think that would help.
Thanks!
Carli
|
By Dave - 4/7/2015
The /skip certainly isn't the problem (even if it applied, it wouldn't skip the entire script, but only the final summary block). Condition assignment *in this script* looks okay, but of course relies on being supplied a numerical groupid. If you were to e.g. pass in a non-numerical ID via URL query parameter, the script would not run. Since you have the logs and other data files available, what are the respective participants' subject and group ids?
> Even if the skip or group assignment items were the issue it doesn't make sense to me why one would skip the IAT script (in
> which both of these are located) and one would skip the demographic question script.
If the demographic script either (1) doesn't need to assign any conditions (because there is only one) or (2) assigns conditions differently (e.g. based on subject number), this can happen.
|
By cfinlm - 4/7/2015
I'm having the participants pick their own 4 digit (numerical) ID codes and the group assignment is random. The ones with issues were S:4192 G:510531161 (IAT skip) and S:1987 G:528070766 (demo skip).
The demographic script doesn't have any assignment conditions.
But actually participant 4192 just emailed back saying she tried another computer (from another network) and I just checked and it didn't skip the IAT script this time which leads me to think it's probably a computer issue on participants' ends.
|
By Dave - 4/7/2015
I agree that those subject and group numbers shouldn't be causing any problem, and I'm fairly sure the cause for the scripts' omission is unrelated to them.
> But actually participant 4192 just emailed back saying she tried another computer (from another network) and I just checked and
> it didn't skip the IAT script this time which leads me to think it's probably a computer issue on participants' ends.
Interesting. Unfortunately I have no spontaneous idea what on the participant's machine could cause such rather weird behavior. If you come across any additional details, please don't hesitate to add them to the thread.
|
By cfinlm - 4/7/2015
Will do! Thanks for your help!
|
|