C# console application to calculate D-scores from IAT data in JSON format. See the sample included for data format/folder organisation. Easily extendable to other input formats (input function to modify is IATBlock.GetQuestions).
See here1 for more information about the IAT, and justification of the scoring algorithm used.
- Extract the IAT "blocks" to be analysed
- For the standard IAT these are blocks 3, 4, 6 and 7.
- Eliminate outliers
- Here outliers are defined as all trials with latencies > 10,000ms.
- Reject subjects
- Subjects who complete more than 10% of the trials in less than 300ms are rejected.
- Calculate mean latency for correct answers
- The mean latency is calculated for the correct responses, to be used in step 6.
- Calculate standard deviation for groups of blocks
- A pooled standard deviation is calculated for blocks 3 & 6, blocks 4 & 7.
- Replace error latencies
- Latencies for incorrect responses are replaced with the quantity mean correct latency (from step 4) + 2 * block SD (from step 5).
- Calculate D-score.
- A D-score is calculated from the resulting data using the formula described in [1].
- Greenwald AG, Nosek BA, Banaji MR (2003). Understanding and using the Implicit Association Test: I. An improved scoring algorithm. J Pers Soc Psychol 85(2):197–216.