Millisecond Forums

Re. #1: Unless the amount of responses outside the interval is excessive, you only drop the individual trials. See Greenwald et al. (2003) for details.
Re. #2: Not necessarily. See Greenwald et al. (2003) and https://www.millisecond.com/forums/Topic3444.aspx

Dear David,

First thank you for your help and support, still I feel a bit lost about the data interpretation.

I calculated the latencies for each participant and also for the whole sample and found the differences between compatible and incompatible.
Still I am not very clear about:

1. Why latencies are transformed in log? (In some studies they dont make the log transform)
2. How we can interpret the error rate? 
3. When dropping latenices outside the elected range, does it affect the result of IAT?
3. Is it possible to create a new variable with the last value of d score and call it "IAT coefficient"?
4. I am not clear about the table of correlation between imolicit and explicit attitudes.

Thank you in advance

> 1. Why latencies are transformed in log? (In some studies they dont make the log transform)

They aren't.

> 2. How we can interpret the error rate? 

Error rates are not commonly interpreted. They are used as a criterion to eliminate entire data sets if excessive.

> 3. When dropping latenices outside the elected range, does it affect the result of IAT?

Greenwald et al. (2003) discusses this in details. The interval was chosen to maximize IAT effects.

> 4. Is it possible to create a new variable with the last value of d score and call it "IAT coefficient"?

The D-score is the "IAT coefficient". If you are using any of the Inquisit 4 IAT scripts, its final value is captured in the summary data file. Also see: https://www.millisecond.com/forums/Topic3444.aspx

> 4. I am not clear about the table of correlation between implicit and explicit attitudes.

I don't know what you are referring to here, but it does not sound like anything related to Inquisit.