In rating experiments analysed with signal detection methods, measures of sensitivity (d(a) and A(z)) and variance (slope of the zROC) are often reported. These pages provide detailed reports of the accuracy (statistical bias) and precision (standard deviation) of each of these summary statistics. See Macmillan, N. A., Rotello, C. M., Miller, J. O. (2004). The sampling distributions of Gaussian ROC statistics. Perception & Psychophysics, 66, 406–421, for discussion of these statistics (Email me for a copy). Please cite the paper and web-site if you use the data herein.
Several variables are included as factors in these tables:
1. NTrials = the number of signal trials on which a subject is tested. Should equal the number of noise trials.
2. NResps = the number of response categories in the rating scale
3. True_da = actual sensitivity, measured as d(a)
4. True_s = actual slope of zROC
5. The symmetry of the decision criteria selected by subjects.
Tables of Symmetric criteria should be used when the area above the upper criterion and the area below the lowest criterion are approximately equal and the other criteria are distribution evenly between them.
Tables of Liberal response criteria should be used when the area above the upper criterion is greater than the area below the lowest criterion (we used a 2:1 ratio in our simulations).
Tables of Conservative response criteria should be used when the area above the upper criterion is less than the area below the lowest criterion (a 1:2 ratio was used.)
5. CritP = A measure of the placement of the most extreme criterion. For Symmetric and Conservative response criteria, CritP measures the area above the most conservative criterion. For Liberal response criteria, CritP (called LowCritP in the tables) measures the area below the lowest criterion.
Go to Tables of Symmetric Response Criteria
Go to Tables of Liberal Response Criteria
Go to Tables of Conservative Response Criteria
email one of the authors:
Neil A. Macmillan Department of Psychology, University of Massachusetts at Amherst
Caren M. Rotello Department of Psychology, University of Massachusetts at AmherstJeff O. Miller Department of Psychology, University of Otago, New Zealand