Hello, I am adapting your 'multi-category' or 'multi-factor' IAT task, and I would like to have the running D score for each of the combinations.. namely ABd, ACd & BCd. I attempted to delete all references to the expression 'D' = that is the category D, not the effect size of d to get three categories rather than 4. But So far I am getting very different results from computing it myself in SPSS.
This is the syntax for calculating the expressions
<expressions>
/ ABm = values.ABsum / values.ABn
/ ACm = values.ACsum / values.ACn
/ BAm = values.BAsum / values.BAn
/ BCm = values.BCsum / values.BCn
/ CAm = values.CAsum / values.CAn
/ CBm = values.CBsum / values.CBn
/ ABsd = sqrt((values.ABss- (values.ABn* (expressions.ABm * expressions.ABm))) / (values.ABn- 1))
/ ACsd = sqrt((values.ACss- (values.ACn* (expressions.ACm * expressions.ACm))) / (values.ACn- 1))
/ BAsd = sqrt((values.BAss- (values.BAn* (expressions.BAm * expressions.BAm))) / (values.BAn- 1))
/ BCsd = sqrt((values.BCss- (values.BCn* (expressions.BCm * expressions.BCm))) / (values.BCn- 1))
/ CAsd = sqrt((values.CAss- (values.CAn* (expressions.CAm * expressions.CAm))) / (values.CAn- 1))
/ CBsd = sqrt((values.CBss- (values.CBn* (expressions.CBm * expressions.CBm))) / (values.CBn- 1))
/ ABBAsd = sqrt((((values.ABn - 1) * (expressions.ABsd * expressions.ABsd) + (values.BAn - 1) * (expressions.BAsd * expressions.BAsd)) + ((values.ABn + values.BAn) * ((expressions.ABm - expressions.BAm) * (expressions.ABm - expressions.BAm)) / 4) ) / (values.ABn + values.BAn - 1) )
/ ACCAsd = sqrt((((values.ACn - 1) * (expressions.ACsd * expressions.ACsd) + (values.CAn - 1) * (expressions.CAsd * expressions.CAsd)) + ((values.ACn + values.CAn) * ((expressions.ACm - expressions.CAm) * (expressions.ACm - expressions.CAm)) / 4) ) / (values.ACn + values.CAn - 1) )
/ BCCBsd = sqrt((((values.BCn - 1) * (expressions.BCsd * expressions.BCsd) + (values.CBn - 1) * (expressions.CBsd * expressions.CBsd)) + ((values.BCn + values.CBn) * ((expressions.BCm - expressions.CBm) * (expressions.BCm - expressions.CBm)) / 4) ) / (values.BCn + values.CBn - 1) )
/ ABd = (BAm - ABm) / expressions.ABBAsd
/ ACd = (CAm - ACm) / expressions.ACCAsd
/ BCd = (CBm - BCm) / expressions.BCCBsd
/ currentblocknumber = expt.1.currentblocknumber
</expressions>
I am computing myself the following on spss based on greewald's (2003) improved scoring algorithm. This is what I'm doing:
1 Computing the "inclusive" standard deviation for all trials in the blocks with the relevant pairings (ie. ABsd, ACsd or BCsd)
2 Computing the mean latency for responses for liking and disliking versions of AB, AC and BC
5 Computing the six mean differences (e.g B liking - A liking , C liking - A liking and so on ... then B disliking - A disliking etc, will produce 6 difference scores )
6 Dividing each difference score by its associated "inclusive" standard deviation
7 This then produces 3 overall Ds (one for each category) which is the equal-weight average of the two ratios for each category
This method is not giving me numbers anything like the above syntax... what am I doing wrong?