Abstract
The Faces test from Wechsler Memory Scale-III is a widely used instrument for detecting impairment in visual memory. However, the degree to which this test actually achieves its purpose may be seriously hampered by error due to guessing. This problem becomes all the more pronounced among elderly clients, where impairment is most likely to be present. Calculations based on the binomial probability function indicate that the probability of scoring in the severely impaired range is extremely low even under a random response pattern, when the norms for older adults are considered. These calculations provide conservative probability estimates that carry direct implications for clinical practice because they apply to the level of the individual client. That is, in clinical practice, detecting severe impairment in older clients, when present, is highly unlikely based on results obtained from the Faces test as currently administered and scored. A revision of the test is warranted.