In many experimental situations, only two response proportions are available: the hit rate (proportion of "old" responses to studied items) and the false-alarm rate (proportion of "old" responses to new items). From those values, we often wish to estimate sensitivity and response bias. A number of estimates of sensitivity have been proposed; we have explored the statistical properties of three of them: d', A z, and A'. A complete description of our study is contained in

Verde, M. F., Macmillan, N. A., & Rotello, C. M. (2006). Measures of sensitivity based on a single hit rate and false-alarm rate: The accuracy, precision, and robustness of d', A z, and A'. Perception & Psychophysics, 68, 643-654.

(email me for a copy)

These pages are intended as a companion to that paper; they include a more complete presentation of the data.

Please cite the paper and the page if you use these data.

 

A table of the statistics is available here, and a description of the column headings in the table is located here.

 

 

For information on the sampling distribution of ROC statistics (including slope) based on rating data, see Macmillan, N. A., Rotello, C. M., Miller, J. O. (2004). The sampling distributions of Gaussian ROC statistics. Perception & Psychophysics, 66, 406–421. (Click here for those web-pages)

Contact one of the authors:

Michael Verde, Dept. of Psychology, University of Plymouth, England

Neil Macmillan, or Caren Rotello, all of the Dept. of Psychology, University of Massachusetts - Amherst

 

back to Caren Rotello's web-page