output of item characteristics into raw data file


Author
Message
JGull
JGull
Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)
Group: Forum Members
Posts: 2, Visits: 8
I'm new to Inquisit. After going through the tutorials and example scripts, I have a good feel for programming an experiment. However, I'd like to know whether it is possible to output item-specific information for each trial into the raw data file. I haven't seen this done in the example scripts I've looked at.

For example, I'm currently preparing a lexical decision task, where participants will respond to words and nonwords. My words and nonwords are controlled on various factors (e.g., word length, frequency of usage, etc.). Optimally, I'd like this item information to be output into the raw data file. So e.g. if on trial 1 the participant sees "cat", then it's associated length and frequency appear in the data file for trial 1. Other software packages I'm familiar (e.g., E-Prime and OpenSesame) can load data from delimited files that contain all the information associated with a trial and it gets output to the raw data file. However, with Inquist, all the examples I've seen limit you to specifying the only the actual stimulus under an item tag.

Is it possible to accomplish this with Inquisit 4? Is this something that's available in Inquisit 5? It's pretty common in Psycholinguistics to analyze item properties and subject properties of the data. I could do a merge of the file that contains item data onto the raw data files after the fact, but this is not ideal especially if the stimuli differ subtly (e.g., contain a space or diacritic get erased in the output; this is more of a problem when considering sentences than individual words).

Thanks!

Dave
Dave
Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)Supreme Being (1M reputation)
Group: Administrators
Posts: 13K, Visits: 104K
Yes, that's absolutely possible. You need to encode the item characteristics somewhere, e.g. in <list> elements, pair those lists to the stimulus (see the "How to present stimulus pairs" topic in the documentation), and then in the trial store the characteristics in global variables (<values> entries) which you log to the data file. In a nutshell:

<values>
/ length = 0
/ frequency = 0
</values>


<block someblock>
/ trials = [1-3 = mytrial]
</block>

<trial mytrial>
/ ontrialend = [values.length = list.length.nextvalue; values.frequency = list.frequency.nextvalue; ]
/ stimulusframes = [1=stimulus]
/ validresponse = (57)
</trial>

<text stimulus>
/ items = ("cat", "bird", "crocodile")
</text>

<list length>
/ items = (3, 4, 9)
/ selectionmode = text.stimulus.currentindex
</list>

<list frequency>
/ items = ("frequency of cat", "frequency of bird", "frequency of crocodile")
/ selectionmode = text.stimulus.currentindex
</list>

<data>
/ columns = [date time subject group blocknum blockcode trialnum trialcode response latency correct stimulusnumber stimulusitem values.length values.frequency]
</data>


JGull
JGull
Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)Associate Member (231 reputation)
Group: Forum Members
Posts: 2, Visits: 8
Okay this makes sense. I will try it tomorrow. Thanks for your reply!

GO

Merge Selected

Merge into selected topic...



Merge into merge target...



Merge into a specific topic ID...




Reading This Topic

Explore
Messages
Mentions
Search