The Wechsler Memory Scale-Revised (Wechsler, 1987) was elaborated by a
dding 20 recognition items for Logical Memory (LM) and by adding recog
nition, matching, and copy trials for Visual Reproductions (VR). Relia
bility was very good for VR copy (alpha=.77, interrater agreement, 100
% within 1 point). Reliability was modest for LM recognition (alpha=.4
2) and for VR recognition (alpha=.58), but reasonable given the small
number of items on each measure. The reliability coefficient was very
weak for VR matching (alpha=.10), limited considerably by the ease of
the task for this healthy sample and by the small number of items. Aft
er statistically correcting for imperfect reliability, validity coeffi
cients (correlations with similar tests that are well standardized) we
re good for all of the new trials (r=.60 to r=1.00). Age- and educatio
n-appropriate reference data are provided using 81 healthy adults. Fur
ther validation using theoretically and clinically meaningful patient
groups would be very valuable. In the meantime, the new trials should
serve the clinician well as screening tools.