Separate questionnaire data-files

[jarvis24]

Ah, that makes much sense. I've actually been thinking of having MediaLab and DirectRT be able to share a responses.xls file. In the meantime, have you tried to save your responses in MediaLab through a responses.xls file (see Advanced Features) and then access those from another application? MediaLab updates--in real time--whatever values you like in there. You may be able to access the spreadsheet values from another program. If you try it with any luck, please let me know!

[Thomas]

Sure.

-Tom

[Thomas]

Hi Blair,
yet, I have not tried to read out the responses.xls from another programm (because this task seems to be a tough nut to crack), but I found another way to get ML responses as stimuli in DRT. I'm(not yet) firm with C++, I used php, webbrower and server.


Ok, here we go (ML/DRT V.2006, I had no time to try the 2008 versions):

I made html-pages (strictly speaking: *.php, it's an html with php-scripts, but it seems that the server needs the php-file ending to parse the php code well) with html-formulas and use php-scripts to write the responses to a separate textfile. Afterwards, I use a short php-page to do the calculations and write the responses to another textfile which is located in the stim-folder of the experiment so that DRT can read it. In ML, I address the pages within the browser "outside" of Medialab, so data is tracked by php and not by ML. Unfortunately, if I use the HMTL-item type ML wants a file and not an url (I guess there is also a thread here about this topic), and php is only parsed if I call the php-files via a web-browser (because, of course, I need a server to parse the php commands, so my php and html-files are located in the htdocs of the server), e.g.via a short pipefile within ML as a command-parameter.
This works very well because now I can "dictate" how the data is recorded and this makes it much easier to read it out afterwards (especially if DRT wants to read it out).
And, to put it a step further, this procedure can not only generate stim-lists for DRT, but also complete DRT-csv-input-files, and I guess this is the "big-party-feature" of this approach (e.g. if you want to do a lexical decision task and want to match control words with unknown stimulus words according to the word length).
I managed it to get the stuff run in three days using the online-tutorials, so, this may also be no problem for anyone other.

Best, Tom

P.S. I forgot to mention that, according to the parameter which you use to write the data-file, the information is may be overwritten. normally I use a short batchfile to copy the datafiles to another folder, or I simply tell my student research assistants who run the experiment to save the files by hand and rename it with the participant-number.

[jarvis24]

Wow, that's some nice problem solving! If you feel inclined please post a super-simple sample that illustrates this in its most basic form. Would be curious to check it out--especially being able to generate a stimlist for directrt from medialab response during the same session--cool stuff! Admittedly, the php and web server requirement may be a tad intimidating for some. I will keep working on a friendly alternative--but until then--great work!

[Thomas]

First of all, thanks for the complement! I really appreciate it! And you are totally right, it requires learning at least basic php-programming and stuff like setting up a server, and normally we all buy experimental-software to get rid of stuff like that ;-). And I really look forward to a more userfriendly click-version of stuff like that.

Anyway, I made a very, very basic example how I did it, but I must confess that we have always done stuff like that with a c++-programm, but I guess it should also work proper with files generated with php. Unfortunately, I have no licence for ML and DRT on my private machine (I do stuff like that in my spare time), so I was not able to test the example, especially if DRT takes the file. But if I call the pages from firefox, the stimlistfile is written in the format in which you also write the normal stimlists by hand. So I hope that DRT has no problem reading the file.
So I hope I have not promised too much, normally everything should work fine.

So, hope it works,
best,
Tom

P.S. You may give me a short reply after testing the stuff?

[jarvis24]

Thanks Tom, much appreciated--I look forward to going through it. Let me know if you come up with any other neat solutions.