With the publication of the new WJ IV, an old score issue has again resurfaced. Examiners have observed that often an individual's overall GIA score is lower than the arithmetic average of the scores for the component tests. This occurs when an individual consistently scores below average on the component tests. For individuals who score significantly above average on most tests, the GIA is higher than the arithmetic average of the components (the opposite effect).
is not a new phenomenon and is NOT unique to the WJ battery. I've written
about this previously with Joel Schneider. A link to a special report can be found here.
I first wrote about this in 1994 and have re-posted this material (from my first WJ
book) for download (PDF file).
these explanations one will see that the “total does not equal the sum
of the parts” phenomenon also occurs on the Wechslers, but it is HIDDEN from view via the fact that the Wechslers use standard scores (mean=10
plus/minus 3) for subtests. In my 1994 explanation, I present a fictitious case where a child
obtains 4's on all WISC-R subtests (-2 SD). Since a standard score
on a subtest of 4 is -2 SD, these scores would be equal to a standard score
(mean=100; SD=15) of 70. The child would have70's for
all subtests. The arithmetic average of all subtests would be 70. So.....is the WISC-R full scale IQ approximately 70? No. It is 59, or 11 points
line. This phenomenon has been around for years and is present on all
IQ tests. It is more obvious on the WJ batteries were all subtests and cluster scores are on a common 100/15 scale. This
is nothing new. If you have been using other batteries (e.g.Wechslers)
you simply have not had the opportunity to observe it.
I am currently working with a colleague on a special Assessment Service Bulletin to explain this score issue. I will notify readers of its availability as soon as it is completed.
Harrison, A. G., Holmes,
A., Silvestri, R., & Armstrong, I. T. (2015). Implications for
Educational Classification and Psychological Diagnoses Using the
Wechsler Adult Intelligence Scale–Fourth Edition With Canadian Versus
American Norms. Journal of Psychoeducational Assessment, 0734282915573723.
Abstract: Building on a recent work of Harrison, Armstrong, Harrison, Iverson and Lange which suggested that Wechsler Adult Intelligence Scale–Fourth Edition (WAIS-IV) scores might systematically overestimate the severity of intellectual impairments if Canadian norms are used, the present study examined differences between Canadian and American derived WAIS-IV scores from 861 postsecondary students attending school across the province of Ontario, Canada. This broader data set confirmed a trend whereby individuals’ raw scores systematically produced lower standardized scores through the use of Canadian as opposed to American norms. The differences do not appear to be due to cultural, educational, or population differences, as participants acted as their own controls. The ramifications of utilizing the different norms were examined with regard to psychoeducational assessments and educational placement decisions particularly with respect to the diagnoses of Learning Disability and Intellectual Disability.
I have not studied the Harrison et al. study in depth, but would like to share, with his permission, portions of an email shared with me by Dr. Larry Weiss, Vice President, Global Research & Development, Pearson Clinical and Talent Assessment (the publisher of the WAIS-IV). Dr. Weiss and I briefly talked about this controversial paper at the recent NASP conference.
Dr. Larry Weiss comments ( email 3-18-15 - shared with permission)
"To follow up on our discussion about the Harrison el al. paper, they found that a large percent of Canadian college students obtained scores below the average range on the WAIS-IV FSIQ when using the Canadian norms. They considered this finding to be highly unexpected for a sample of college students, and questioned the validity of the WAIS-IV Canadian norms. However, the authors of that study did not adequately take into account that 75% of their sample had clinical diagnoses.
To demonstrate the impact of clinical status on IQ test scores, my research team drew a sample of American subjects matched to the Harrison sample on clinical status and educational level. We then scored the American sample on U.S. norms, and found that the percentage obtaining below average FSIQ scores was almost identical to that reported by Harrison et. al. using Canadian norms. This demonstrates that the Harrison et al. findings are not unique to the Canadian norms, but are due to the mixed clinical status of their sample. Details of our matched sample analysis will appear in an upcoming issue of the Journal of Psychoeducational Assessment in an article by Miller, Weiss, Beal, Saklofske, Zhu, & Holdnack."
Although I will not comment on the specific WAIS-IV study and methodology in question, I can point out that in a special ASB (ASB #12 Use of the Woodcock-Johnson III NU Tests of Cognitive Abilities and Tests of Achievement with Canadian Populations) where I and others reported the performance of a sample of Canadians on the WJ III NU battery, when the Canadian subjects (who were a randomly selected representative sample--not a largely clinically preselected sample), where matched on critical demographic variables to a sample of US subjects, we found that "while some minor score differences are reported across the two samples,
the study findings generally support the use of the U.S.-based WJ III NU
norms with Canadian school-age populations." In other words, when comparable (demographically matched) Canadian and US subjects were compared on the WJ III NU cognitive battery, no significant Canadian-US IQ scores, beyond some minor exceptions, were found.
Readers should wait until the Miller et al. (in press) response paper is released before jumping to any quick conclusions.
[Conflict of interest disclosure: I am a coauthor of the WJ III and WJ IV, a direct competitor to the Wechsler batteries]
Later-born cohorts of older adults tend to outperform earlier born on fluid cognition (i.e., Flynn effect) when measured at the same chronological ages. We investigated cohort differences in level of performance and rate of change across three population-based samples born in 1901, 1906, and 1930, drawn from the Gerontological and Geriatric Population Studies in Gothenburg, Sweden (H70), and measured on tests of logical reasoning and spatial ability at ages 70, 75, and 79 years. Estimates from multiple-group latent growth curve models (LGCM) revealed, in line with previous studies, substantial differences in level of performance where later-born cohorts outperformed earlier born cohorts. Somewhat surprisingly, later-born cohorts showed, on average, a steeper decline than the earlier-born cohort. Gender and education only partially accounted for observed cohort trends. Men outperformed women in the 1906 and 1930 cohorts but no difference was found in the 1901 cohort. More years of education was associated with improved performance in all three cohorts. Our findings confirm the presence of birth cohort effects also in old age but indicate a faster rate of decline in later-born samples. Potential explanations for these findings are discussed. (PsycINFO Database Record (c) 2015 APA, all rights reserved) ----
A general factor of intelligence fails to account for changes in tests' scores after cognitive practice: A longitudinal multi-group latent-variable study
Estrada, Eduardo; Ferrer, Emilio; Abad, Francisco J.; Román, Francisco J.; Colom, Roberto
Intelligence, Vol. 50 – 2015: 93 - 99
I am pleased to announce that version 1.1 of the WJ IV Evolving Web of Knowledge (WJ IV EWOK) is now available. This is your one stop internet source for keeping abreast of announcements, research, etc. regarding the new Woodcock-Johnson IV Battery. It is the same concept used for the prior WJ III EWOK.