Millisecond Forums

sign of D score

https://forums.millisecond.com/Topic6711.aspx

By adkatz2 - 10/5/2011

Thanks
in advance for helping me with this question.  I'm embarrassed to say 
that I'm having trouble interpreting the D-score based on my reading of
Greenwald's SPSS scoring syntax, so I want to check my understanding. 
His syntax says that:


". . . high scores indicate stronger association of 'attribute_left_label'
with 'TARGET_2_LABEL' than 'attribute_right_label' with 'TARGET_2_LABEL.'"


I used the racial attitudes IAT program provided by Greenwald at his website in my data collection, but I can't determine how these instructions correlate with "attribute_right_label" and "TARGET_2_LABEL."  If I'm reading the literature correctly, however, it would appear that higher D-scores demonstrate a preference for compatible over incompatible associations--is this correct?  So, since "White" is typically associated with "pleasant" for most White Americans--and is therefore compatible--a higher D-score for a White participant would indicate greater compatible associations (higher pro-White bias and greater automatic prejudice toward Blacks).  On the other hand, White Americans with lower automatic prejudice toward Blacks (or Blacks with pro-Black bias) might have low or even negative D-scores.


I think I ran the scoring syntax correctly, but I have a mix of negative and positive scores, so I want to make sure I understand who is who.  I hope this question is clear--Thanks!

By Dave - 10/5/2011


His syntax says that:


". . .

Normal
0


false
false
false







MicrosoftInternetExplorer4



high scores indicate stronger association of 'attribute_left_label'
with 'TARGET_2_LABEL' than 'attribute_right_label' with 'TARGET_2_LABEL.'"



Interpretation will obviously depend on which actual concepts attribute_left_label and target_2_label, etc. represent in your actual script. Just look it up! E.g. does target_2_label mean "white" in your script? Or is it "black"? And so forth...


If I'm reading the literature correctly, however, it would appear that higher D-scores demonstrate a preference for compatible over incompatible associations--is this correct?  So, since "White" is typically associated with "pleasant" for most White Americans--and is therefore compatible--a higher D-score for a White participant would indicate greater compatible associations (higher pro-White bias and greater automatic prejudice toward Blacks).  On the other hand, White Americans with lower automatic prejudice toward Blacks (or Blacks with pro-Black bias) might have low or even negative D-scores.


Yes, that is correct.


Regards,


~Dave


By adkatz2 - 10/6/2011

Thanks for confirming my intuition.


I tried looking it up, but I was thrown by "TARGET_2_LABEL."  I mean, I can figure out that "attribute_right_label" is "pleasant" in the race IAT, because it is always positioned on the right and only the "Black" and "White" labels shift around.  But I couldn't find anywhere in the experiment file where the targets are numbered 1 or 2. . . .


In any case, I think I know what happened and hopefully the data scoring will be simple to finish.  Take care!