How to interpret IAT scores in Inquisit


Author
Message
Dave
Dave
Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)
Group: Administrators
Posts: 11K, Visits: 68K
#1: Look at the script's <expt> and/or <variables> elements. If you've used any of the Inquisit *4* IAT scripts, /groupassignment will have been set to 'groupnumber' *not* 'subjectnumber*, i.e. block order is determined by the groupid *not* subjectid.

If this is the case, then

IF (MOD(subject,2) = 1) ORDER = 1 .
IF (MOD(subject,2) = 0) ORDER = 2 .

is obviously not the right way to determine block order. The group number has to be used instead.

#2: If those are valid paths on your computer and you do have write permissions for this location (which you should): Yes.

Ana_Oliva
Ana_Oliva
Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)
Group: Forum Members
Posts: 10, Visits: 33
Thank you again Dave.

I am using Inquisit 3... Here's what I found in my IAT script:

***********************************************************************
Experiment
***********************************************************************

<defaults>
/ fontstyle = ("Arial", 3.5%)
/ screencolor = (0,0,0)
/ txbgcolor = (0,0,0)
/ txcolor = (255, 255, 255)
/ minimumversion = "3.0.0.0"
</defaults>

<expt>
/ blocks = [1=block1; 2=attributepractice; 3=block3; 4=block4; 5=block5; 6=block6; 7=block7; 8=block8; 9=block9]
</expt>

<variables>
/ group = (1 of 2) (block1=targetcompatiblepractice; block3=compatibletest1; block4=compatibletestinstructions; block5=compatibletest2; block6=targetincompatiblepractice; block7=incompatibletest1; block8=incompatibletestinstructions; block9=incompatibletest2)
/ group = (2 of 2) (block1=targetincompatiblepractice; block3=incompatibletest1; block4=incompatibletestinstructions; block5=incompatibletest2; block6=targetcompatiblepractice; block7=compatibletest1; block8=compatibletestinstructions; block9=compatibletest2)
</variables>

So, I guess that this means that I am not using the correct syntax...

How (and where) can I correct the syntax in order for it to make the proper calculations then?

Thank you!

Ana
Dave
Dave
Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)
Group: Administrators
Posts: 11K, Visits: 68K
No, you don't have to correct or modify the SPSS syntax -- Inquisit 3  IAT scripts assign block order based on subjectnumber just as the SPSS syntax assumes.

Ana_Oliva
Ana_Oliva
Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)
Group: Forum Members
Posts: 10, Visits: 33
No? :D What a relief! :p

So, I interpreted the results in the right directions and don't need to recode anything? :D I've never been so happy for not making an upgrade...! :p

I'll try the alterations regarding the errors then... I hope that it all goes well now.


Thank you! :)
Ana_Oliva
Ana_Oliva
Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)Esteemed Member (1.9K reputation)
Group: Forum Members
Posts: 10, Visits: 33
I have already introduced the alterations and all is fine!!! Not one single error!

Thank you, you're a life saviour! :D


Ana
Kala
Kala
Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)
Group: Forum Members
Posts: 10, Visits: 22
Hello,

I have questions about two general topics. First, is the interpretation of the BIAT D score the same as the IAT score? Sriram & Greenwald (2009) explain how to interpret D scores above zero, but as far as I can see do not explicitly state how to interpret scores below zero (i.e., negative scores). I have negative score means in my data, so I want to be absolutely sure of my interpretation.

Second, I am curious how the "expressions" variables are calculated by the Inquisit software. Does the running tally of D take the order of the trials into account? Also, does the "DInquisit" variable that is generated in the "Criteria" outfile step of the SPSS script a reliable metric for analysis? I'm unsure what the "LAST" command in the syntax file refers to. I ask because when I use the D generated by the syntax file I get negative score means, but when I use the DInquisit I get positive score means.

Thank you for your assistance!

Dave
Dave
Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)
Group: Administrators
Posts: 11K, Visits: 68K
As detailed in the 1st post of the thread, D-scores can be either positive or negative. The sign indicates the "direction" of the effect. It makes no difference in this regard whether you're looking at D-scores stemming from a Brief IAT or from a "standard" IAT.

> Does the running tally of D take the order of the trials into account?

The order of trials does not matter. D is a difference score, in essence (mean latency in incompatible condition - mean latency in compatible condition) / normalized by standard deviation. Each line in the raw data file can be read as "this is the D-score *given all the data available up to this point*.

> Also, does the "DInquisit" variable that is generated in the "Criteria" outfile step of the SPSS script a reliable metric for
> analysis?

That's simply the D-score *as calculated by the script and recorded to the data file*. The LAST function extracts the value from the last line, i.e. the final D-score (when all relevant data are in). For general guidance on SPSS syntax, please the documentation provided by IBM / SPSS.

Kala
Kala
Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)
Group: Forum Members
Posts: 10, Visits: 22
Thank you for the quick reply! At this point I am trying to figure out why, when I run the SPSS syntax file, the D scores are negative by the DInquisit scores are positive. It seems that although they may not be precisely the same number (because of the 300 ms latency correction in the calculation of the D score), they should at least be the same direction most of the time.

Dave
Dave
Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)Supreme Being (930K reputation)
Group: Administrators
Posts: 11K, Visits: 68K
Make sure you use the syntax correctly and adapt it as needed. Particularly check how your *script* assigns between-subjects conditions (Based on subject id? Based on group id? Randomly?) and verify that the SPSS syntax reflects that.

Kala
Kala
Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)Esteemed Member (2K reputation)
Group: Forum Members
Posts: 10, Visits: 22
Thank you. I am going through the syntax file and my Inquisit .exp file with a fine-tooth comb. Just in case you can offer some insight, I have posted how I assigned BIAT order, as well as the relevant section of SPSS script that decodes the order, pairing, and test conditions, below. I adapted the syntax file by putting in the correct blocknums for my BIAT blocks in the VALUE LABELS section and in the COMPUTE Test section; other than that I did not see where I needed to make additional adaptations.

<expt>
/ subjects = (1 of 2)
/ blocks = [1=BSRI;2=READ;3=ManipCheck;4=IATinstructions;5=short_a;6=short_b;7=A;8=B;9=A;10=B;11=ROMANTIC;12=CQ;13=opentext2;14=demographics]
/onexptend = [values.completed = 1]
</expt>

<expt>
/ subjects = (2 of 2)
/ blocks = [1=BSRI;2=READ;3=ManipCheck;4=IATinstructions;5=short_b;6=short_a;7=B;8=A;9=B;10=A;11=ROMANTIC;12=CQ;13=opentext2;14=demographics]
/onexptend = [values.completed = 1]
</expt>

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

GET TRANSLATE FILE = 'BriefIAT2.dat'
  /TYPE=TAB /MAP /FIELDNAMES .

VALUE LABELS blocknum
 5 'Attribute practice'
 6 'Attribute practice'
 7 'First pairing practice'
 8 'Reverse pairing practice'
 9 'First pairing test'
 10 'Reverse pairing test'.

IF (MOD(subject,2) = 1) Order = 1 .
IF (MOD(subject,2) = 0) Order = 2 .

COMPUTE Pairing = 0.
IF (blockcode = 'A' ) Pairing = 1.
IF (blockcode = 'B' ) Pairing = 2.

COMPUTE Test = 0.
IF (blocknum=7|blocknum=8) Test = 1.
IF (blocknum=9|blocknum=10) Test = 2.
GO

Merge Selected

Merge into selected topic...



Merge into merge target...



Merge into a specific topic ID...




Reading This Topic

Explore
Messages
Mentions
Search