Millisecond Forums » Millisecond Forums » Inquisit 5 » D value for compatible and not compatible blocks

D value for compatible and not compatible blocks

missnatalien — Fri, 07 Jun 2019 09:16:03 GMT

Hi,
I have a few questions regarding the D value. My supervisor would like me to count the D value for the compatible block (in my case those are blocks 3 and 5) and for the incompatible blocks (in my case 7 and 9). Now, obviously, the D value for the compatible blocks is counted - this is the value in the column expressions.d in the last row of the last compatible block. However, it is not the same case for the incompatible blocks - this one I have to count by myself, right? Because the last D value of incompatible blocks is the actual D value (so the actual IAT effect).

Now, I have a question regarding interpreting the data. If the D value for the compatible blocks equals -1,4592282838115384624, then what does that tell me? Does it mean it was easier for the person to put the positive words to their own group? When is the score negative and when is it positive? How would that look for the incompatible blocks?

For the record, in my IAT it looks like this: attributes A are positive, attributes B are negative, target A is their own group, target B is the other.

Thank you in advance for the help!

RE: D value for compatible and not compatible blocks

Dave — Fri, 07 Jun 2019 09:16:03 GMT

[quote]

[b]missnatalien - 6/7/2019[/b]

[quote]

[b]Dave - 6/7/2019[/b]

[quote]

[b]missnatalien - 6/7/2019[/b]

[/quote]

You're misunderstanding D. There can be no such thing as a D value for the compatible blocks and a D value for the incompatible blocks. The D-score does -- and must by definition -- reflect both. Simplified, D is calculated as

D = (mean response time in incompatible blocks - mean response time in compatible blocks) / standard deviation of response time in all blocks

There is nothing to count. The D value in the final row of your data set is the final and only D value. All values before that are merely "this is the D score based on the data up to and including the current trial."

[/quote]

Okay, thank you very much. I think I should talk to my professor now and explain the same thing to him. He wanted me to compare compatible blocks and incompatible blocks. How should I do that then? I have two samples: 40 Germans and 40 Poles. He wanted me to compare the compatible blocks between Germans and Poles and to do the same with incompatible blocks. Is that possible then?

[/quote]

Not with D-scores, no. As explained above, D-scores are difference scores (difference in performance in incompatible vs compatible blocks).

To expand: A D score already compares compatible vs incompatible.

Now, let's say you conducted a Germany vs Poland IAT. What you'd expect, then, is that your German participants' D scores would reflect preference for Germany (their in-group), whereas your Polish participants' D scores would reflect a preference for Poland (their in-group). And that's what you'd want to compare in a scenario like this.

RE: D value for compatible and not compatible blocks

missnatalien — Fri, 07 Jun 2019 09:12:24 GMT

[quote]

[b]Dave - 6/7/2019[/b]

[quote]

[b]missnatalien - 6/7/2019[/b]

[/quote]

RE: D value for compatible and not compatible blocks

Dave — Fri, 07 Jun 2019 09:05:49 GMT

[quote]

[b]missnatalien - 6/7/2019[/b]

[/quote]