[R] Calculate Az (A sub z) with R?

Hans Werner Borchers hwborchers at googlemail.com
Fri Jul 15 23:05:58 CEST 2011

dd2es <dd2es <at> virginia.edu> writes:

> I am looking for (or interested in writing) a function that calculates Az,
> an alternative measure of discriminability from SDT (alternative to d', Az).
> I have written my own functions for d', A', B"d, and am aware of the 'sdtalt'
> package, but I have yet to find a way to calculate Az, since it require the
> phi operator.

The Phi --- not Psi, see the paper --- function is simply the cumulative
normal distribution, so you can use pnorm() instead.

--  Hans Werner

> For a relevant paper/discussion (and formula), please see Verde and
> McMillian, 2006 (Measures of sensitivity based on a single hit rate and
> false alarm rate: The accuracy, precision, and robustness of d', Az, and Az)
> Any help on this would be greatly appreciated!
> David Dobolyi
> Graduate Student
> Cognitive Psychology
> University of Virginia

More information about the R-help mailing list