Concerns about the WAIS 4 and WMS 4

Loring DW, Bauer RM.  Cautions and concerns regarding the newWechsler IQ and Memory Scales.  Neurology 74: 685-690 2010.

The new tests have fundamental changes from the old versions and are not established in various populations that they intend to serve.  Although the guidelines require neuropsychologists to use the "most current" tests, Bauer and Loring argue that the old tests are better validated and should be used preferentially in many cases.They compare the new versions of the WMS and WAIS to Windows Vista software release-- those who wish to run Windows XP should not be prevented from doing so.

Authors note the WAIS traditional scores of VIQ and PIQ are not truly measures of verbal and nonverbal intelligence as originally conceived, but cluster into 4 domains of verbal comprehension, perceptual reasoning, working memory, and processing speed. VIQ contains measures of verbal abstraction (Similarities) and knowledge (Information), also attention and working memory (Digit Span).  PIQ shows visuospatial problem solving (Block Design) and processing speed (Digit Symbol).  In WAIS III, 4 factor composites could be calculated, along with PIQ and VIQ. 

In WAIS 4, only FSIQ can be calculated, with VIQ and PIQ eliminated.  VCI and PCI are "substituted" for VIQ and PIQ for "decision making."  VCI and PCI exclude working memory and processing speed, which account for 40 % of the variance of FSIQ.  The General Abilities Index (GAI) is introduced, which also excludes working memory and processing speed.  GAI does not replace FSIQ, and VCI and PCI do not measure the same things as VIQ and PIQ.

The WAIS subsets have been changed to decrease emphasis on speeded tests, so that they are less sensitive to diseases that show psychomotor slowing (such as Parkinson's disease) and slowed information processing (such as multiple sclerosis).  The new WAIS will most likely have less patients with FSIQ under 70, leading to epidemiologic mismatches and skews in historical studies and less people qualifying for disabilities.  This is because many diseases with disability depend on the measurement of psychomotor slowing. The ostensible reason for the new version, to recalibrate the average IQ to 100, may not be met since the amount of change in the FSIQ mean is less than the variance of the test.

The WMS-4 incorporates changes, but many previous changes have fallen flat and then been eliminated in the "next" edition of the test.  The most important change since 1945 is the inclusion of the 30 minute delayed recall of the Logical Memory and Visual Reproduction subtests. In the WMS 3, previous concerns that were addressed failed to hold up under clinical experience and subsequently had to be dropped, analagous to a new software release.  In the WMS-4, the second trial for repetition of one of the to Logical Memory stories is inexplicably dropped, even though one of the reasons for the failure to find material specific memory impairments is the use of single trial memory tasks instead of learning over trials.  New nonverbal memory test within a grid may prove beneficial, because it requires remembering both visual and spatail information, but that has yet to be determined empirically.  Moreover, validity of the tests to specific diseases and prediction of clinical outcome is absent with the new tests.  For example, only 8 patients who had undergone temporal lobectomy were in the sample.  In some clinical trials, the HVLT or RVLT have been chosen because they are unlikely to be revised.  Test retest data also does not exist, and will not for many years, with the new forms of the tests, because data was not obtained prior to publication. 

Authors conclude that the old tests are not outdated, but proven useful, and that the new versions of the old tests do not deserve to be considered automatically current without further validation.