In Reply.
—STATPAC does indeed provide a plethora of quantitative data. We chose to concentrate on the summary scores, pattern standard deviation (PSD), and mean deviation (MD) because these provide the practicing ophthalmologist and screening program director with comprehensible indexes rooted in probability theory. Our major question was whether these easily grasped figures did a reasonable job in distinguishing normal subjects from those likely to have glaucomatous optic nerve damage. Our assessment was that PSD did, particularly when the subject matched the "reliability" profile of those in STATPAC's normal database.We have yet to complete similar analyses for STATPAC probability plots. These are far more difficult to approach in a rigorous, quantitative fashion. Like Heijl et al, we have encountered one or more "abnormal" points in these plots when the global indexes have been normal. But this has occurred in entirely normal subjects as well as in those suspected to