|
PiotrW
|
|
|
Group: Forum Members
Posts: 9,
Visits: 13
|
Thanks for all your help, Dave
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
Data output for <survey> elements cannot be customized. The <data> element has no effect on <survey> data and isn't supposed to (I am sure about that).
|
|
|
|
|
PiotrW
|
|
|
Group: Forum Members
Posts: 9,
Visits: 13
|
I was hoping to use <data> to modify what is written to the default data file in the last script but it fails. Is there any way to customise the contents of that default results file in a "pure survey" other than - the dummy experiment you suggested?
(I am grateful for that, as it works, but is not quite self-explanatory and I might struggle to understand what the lines were intended to do some year or so ahead)
|
|
|
|
|
PiotrW
|
|
|
Group: Forum Members
Posts: 9,
Visits: 13
|
Dave, my mistake - I did not notice the creation date of my summary file. It too, was not being updated since upgrading the inquisit.
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
Re. #1: As I said, I'm not sure about that. I'll have to discuss with the developers, what they intended <summarydata> to do (or not to do) in a situation such as yours. BTW, the file you attached does not produce a summary file for me.
Re. #2: Yes, that would have been 4.0.2.0 or earlier, I believe. Again, not sure whether this was ever the intended behavior or not (cf. #1).
I'll make sure this gets clarified one way or the other.
|
|
|
|
|
PiotrW
|
|
|
Group: Forum Members
Posts: 9,
Visits: 13
|
Dave Many thanks for the workaround!
1) Do you mean that even for questionnaires a dummy expt / block / trial structure should be used as a rule? I attach one that has no problem writing to a summary even though it uses only survey command. Should I modify it too to be safe? 2) Re: bug - the very same script was running ok. on an earlier version of Inquisit.
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
Thanks for the files. I am frankly not sure if what you are seeing is a bug or not (will have to take up discussion about this with the developers). With that caveat, It's always been my understanding that <summarydata> is supposed to work when there is a traditional <block> / <trial> structure, not when solely <survey> elements are used (whose data output cannot be customized anyway). In any case, there is a straightforward way, to get the data you want. Taking 'DichoticDRM_test.iqx' as example, you can add
<expt> / blocks = [1=DichoticDRM_test; 2=summaryblock] </expt> <block summaryblock> / trials = [1=summarytrial] </block> <trial summarytrial> / trialduration = 0 / validresponse = (noresponse) </trial>
to guarantee the creation of a summary data file.
|
|
|
|
|
PiotrW
|
|
|
Group: Forum Members
Posts: 9,
Visits: 13
|
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
> The problematic scripts are too long to include here, but of course I can send them if necessary.
You can *attach* files to a thread when posting a reply, so script length is not an issue. Please do so. Thanks.
|
|
|
|
|
PiotrW
|
|
|
Group: Forum Members
Posts: 9,
Visits: 13
|
Hello,
We were running several scripts succesfully on Inquisit v. 4.0.2 in May 2013. Now we decided to continue our research and ran the very same scripts again on v. 4.0.4 and 4.0.5. Only after completing this part of the study we realized, that two important summary data files are dated May 2013, and contain only old data. No new data was added to them. This was hard to spot as all raw files and summary data files in other scripts were ok. Unfortunately, in effect, we lost many hours of testing...
At first we thought we somehow accidentally overwritten the new summary files with the old ones even though it seemed quite unlikely. But I inspected the problem closer and run the scripts in an empty folder, where there was no summary data file to append to. None was created. Only raw data. The original experiment was run on PCs, I am now testing the script on Mac. Same problem on both platforms.
I read the change log of Inquisit, and if I got it right, some bug was corrected that made the program fail to write summary data while writing the raw data ok. Just our problem. Looks like the correction did solve it in some cases but introduced it in other.
The problematic scripts are too long to include here, but of course I can send them if necessary.
|
|
|
|