First page
Back Continue Last page Summary Graphics
We are currently running a psychoacoustic experiment where we have looked at this problem of common noise sampling variability.
We used three stimulus sets of size 500 and therefore ran three experiments. The results can be pooled give 750 stimuli per event or they can be analysed separately to give an estimate of the amount of variability for sample sizes of 250 per event.
Shown are three 250 sample per event GOC curves for myself and one 750 sample per event. Each GOC curve is based on 16 replications of the experiment, the 16 ROC curves are those below the GOC curves. The ROC curve at the top is the presumed theoretical.