Jens,
This doesn’t work for my case, but it did give me an idea that worked!
I have the same number of trials (40) in every block, so I created a variable where the new reaction time was added after each trial, and then the average was calculated by dividing it by the number of trials.
##Begin Experiment
#rt variable
rt_trial = 0
##End Routine
rt_trial += key_resp_5.rt
rtMean = rt_trial/40
newTime = rtMean*multiNum