Sizes of stimulus sets in the literature

A brief overview of the literature shows people using stimulus sets of around 20 stimuli per event. Replicating such experiments with a different stimulus set could result in near perfect performance, or near chance performance simply due to sampling variability.

I do realise that it sometimes impracticle, if not impossible to have larger sets of stimuli, but the researcher needs to proceed with caution when interpreting such results.

In sum, measuring common noise sampling variability gives us a way of deterimining the minimum sized stimulus set in an experiment. It is useful because it also gives us an estimate of the variability in the resulting measures of observer performance.