Hey,

So I'm including an IAT into my survey but don't really want my participants to know what their IAT score is. I tried deleting the 'Summary block'  but that resulted in the loss of the expression.da, expression.db and expression.d from the data output file. 

Any thoughts?

Deleting the summary block (or rather removing it from your <expt> / <variables> elements so it's not run) should in no way, shape or form affect data recording. The summary block is not instrumental in that regard. You may have accidentally deleted or modified other parts of the script (e.g. the <data> element). I'd recommend you go back to the unmodified original and compare it to your current script to get an overview of the differences.

Regards,

~Dave

That is what I thought.......

I did go back and compare it to the original IAT script....I even tried reinserting the "summary block" but still have not got anything. 

Here is my data columns section:

<data>

/ columns = [date, time, subject, blockcode, blocknum, trialcode, trialnum, response, correct, latency, stimulusnumber, stimulusitem, expressions.da, expressions.db, expressions.d ]

</data>

I wont bother posting each block, but it is identical to the original.

Where should  expression.db etc be calculated?

The various D measures are running calculations, i.e., their values will be updated on a trial-by-trial basis as the relevant data trickles in. As long as no such data is available (e.g., at the beginning of the task during attribute practice), there will of course be no data in the respective columns -- no data, no Ds.

Generally see https://www.millisecond.com/forums/Topic3444.aspx.

The link you provided was broken.....

And Im not too clear on what you meant by the D measures being running calculations......I have been testing the data repeatedly, but the dat files still dont show anything

The URL works fine for me. I don't know why it wouldn't for you, but you might as well search the forum for "How to interpret IAT scores".

It means the scores are updated on a trial-by-trial basis as more and more relevant data become available. Thus the last line of data for a given subject will hold the final scores.

Then please provide both the script as well as an example data file demonstrating the issue (you may put them in a zip file and attach the archive to this thread via the 'Options' tab when posting a reply). Please also provide the exact Inquisit version you are using. Without any of this info, neither I nor anybody else will be able to tell you what might be amiss.

Whoops.....I didn't delete the full stop at the end of the url. 

But I had already read this amongst the other pages provided there.

[View:http://www.millisecond.com/community/cfs-file.ashx/__key/CommunityServer.Discussions.Components.Files/14/6644.Pride-Experiment.zip:550:0]

Thats the files you requested. Thanks for your help

The data file you attached clearly has expressions.da, expressions.db, etc. recorded in it. While the records for subject #1 look a bit suspicious at first glance (no data in expressions.da), the records for subject #4 look as expected.

Given the above (the expressions *are* logged to the data file), I admittedly don't understand your question / problem description.

Hey, yeah it seems to work almost randomly......for some subjects (in either conditions), it comes up perfectly but then it doesn't for others. 

I'm assuming its fine if the D score doesn't appear until the very last row, but in some instances it is not appearing until a completely different trial. 

Also....worse comes to worst, would I be able to calculate the D-score myself in SPSS later using the response? I ask because I have already put the experiment online and have people responding now. So if this isnt the case I will be forced to delete the data of all those who are responding now (until I fix this issue). 

Thanks

Weird. I have no idea why that would be -- might be a bug in the Inquisit 4 beta (if that's what you are using), in which case you should report it to inquisit@millisecond.com. Perhaps this only happens on Macs or something like that...

Certainly reason to be suspicious and check your data / re-calculate D from the raw data in your preferred stats package.

Yes. You can either write your own syntax based on a solid understanding of the necessary calculations involved in determining D (cf. Greenwald et al., 2003) or you can start from the SPSS syntax files available via the IAT download page at millisecond.com. You'll have to modify the syntax to accommodate for the various changes in your script (additional blocks, thus different block numbers, etc.).

sigh....... 

Guess I'm going to have a whole lot of work ahead of me to calculate the raw data.......that is seriously disappointing.

Do you think it is worth me rewriting the whole thing in version 3 then? 

It's not that hard, really.

If you're not using any new functionality specific to Inquisit 4 (you probably don't), the existing script should not require any major modifications (if any at all).

Ok, last question.....

Do you think it's easier to reupload my script online as Inquisit 3 (assuming this is what's causing the problem) or to just calculate the data later (in the absence of the D score)? 

Keeping in mind that my experiment is already up and running online.

Thanks for all your help

Personally, I would opt for re-calculating scores based on the raw data. And, again personally, I would do this anyway to check for consistency, screen for outliers, etc. -- I've always felt that getting to know your data from the ground up is a valuable effort in and of itself.