Pearson: The original Kaufman Assessment Battery for Children was published in 1983. What factors led to the revision?
Dr. Kaufman: Any test has to be revised — children get measurably smarter every decade, so test norms must be refreshed periodically. Our goals were to create a test that's more current and more responsive to recent research and theory. We wanted to be sensitive to the feedback we received about the original K-ABC — both pro and con — and make some key changes during the revision process to respond to today's environment.
The K-ABC was the first ability test to include a theoretical base. Others have built on our original work, and have developed tests based on several contemporary theories. We wanted to do the same to enhance and strengthen the theoretical base of the KABC-II to make it more contemporary. We wanted our revision to reflect current theories in neuropsychology and cognitive psychology. And, as we've always tried to do, we built the new KABC-II to effect change and help children.
Pearson: Please talk about the process of developing the new test—what were your concerns and priorities as a test developer? How were they addressed?
Dr. Kaufman: The first K-ABC measured two areas of ability — Sequential Processing and Simultaneous Processing. For the KABC-II, we wanted to include a wider array of abilities, including practical abilities. Examining areas like learning ability and planning ability permits us to help children better by giving us an indication of how they think and reason, how they employ higher level skills and how they use cognitive processes. We also extended the age range of the KABC-II to cover all ages between preschool and high school (3-18 years). The original K-ABC was normed between ages 2:6 and 12:6.
Pearson: What were your primary goals in the test development, and how did you go about accomplishing them?
Dr. Kaufman: For this test revision, we decided to blend two different theories. As with the original test, we used the Luria model, based on mental processing abilities. The KABC-II also includes the Cattell-Horn-Caroll (CHC) theory, which is well known to psychologists and in test construction. It uses a Gf/Gc, or fluid and crystallized ability model, to understand cognitive abilities.
We wanted to include both of these models-one more neuropsychological (Luria) and one more psychometric (CHC) — to ensure a substantial theoretical base. It helps us reach our practical goal of creating a broad measure that offers parents and teachers good information about a child's ability.
Pearson: What should psychologists who are familiar with the original K-ABC expect from the new test? What are the differences and similarities?
Dr. Kaufman: There were several things that we felt were very important to include in this revision. The K-ABC showed reduced differences in test scores between groups of white and minority children relative to differences found on traditional IQ tests. In other words, it was a fairer measure of the mental abilities of children from culturally different backgrounds.
The original K-ABC separated mental processing and achievement. This structure was both praised and criticized. It promoted fairness for children of non-mainstream or bilingual backgrounds, because the mental processing tasks were not heavily dependent on factual knowledge and de-emphasized verbal responses.
However, factual knowledge such as vocabulary and language ability are aspects of cognitive ability, especially for those children who have been given appropriate opportunities to learn-gifted children, as an example. So, the same factors that made the K-ABC more fair for some children were criticized as being less fair for gifted or highly verbal children.
As a result, we've added a second global score to the KABC-II to address this issue. The Luria model resembles the original test — its global score is called the Mental Processing Index (MPI) and it emphasizes mental processing (how children solve problems). The scale that measures language abilities and word knowledge (Knowledge/Gc) is excluded from the MPI. The new CHC model includes the Knowledge/Gc scale as a fundamental component of its global score, the Fluid Crystallized Index or FCI.
This dual model requires the examiner to make a choice, and we've provided guidelines to help examiners select the most appropriate model for each child they evaluate. Under ordinary circumstances, the CHC model should probably be chosen. However, if the examiner thinks a child might be unfairly penalized by the inclusion of the Knowledge/Gc scale, or if the examiner believes that knowledge is not a fair measure of a child's ability, then the Luria model and the Mental Processing Index (MPI) should be used. For example, it is usually a good idea to select the Luria model for children who are bilingual, who are referred for an expressive or receptive language disorder, or who are from non-mainstream backgrounds. (Of course, the selection of model must be made before the KABC-II is administered, not after.)
In general, the choice of model is based on the child's background and the reason for referral. There is, however, one exception. If the examiner prefers the Luria model (for example, a long-time K-ABC user who prefers to keep mental processing ability separate from language abilities and knowledge), she or he can choose to use the MPI, regardless of the reason for referral.
Pearson: What aspects of the K-ABC were kept and why?
Dr. Kaufman: The K-ABC included several innovations. It was the first test to include teaching items to help children understand the tasks they were being asked to perform. Again, this was a fairer way to assess each child referred for evaluation. The teaching items helped ensure that the child's ability to perform a task was measured — not their ability to understand instructions. This practice has been retained for the KABC-II, and new editions of other tests have followed suit.
We also published a thorough manual to accompany the original K-ABC, a manual that included the results of more than 40 validity studies along with specific guidelines for test interpretation. We began this approach with the K-ABC, and other tests have again followed suit. Like the original K-ABC, the KABC-II includes a thorough, informative manual.
Pearson: How do the standardization and norm samples reflect on the test's methods for assessing ability, especially with children of diverse cultures? Can you describe some of the validity studies that are now underway?
Dr. Kaufman: All norm samples for all tests must be fair — that is, they must include all SES (socio-economic status) groups, all ethnic groups, and students with special needs, all in the correct proportions to match the US Census. All tests do it, it's essential.
Validity studies are conducted to show that a test measures what it's supposed to measure. Confirmatory Factor Analysis, or CFA, is a sophisticated statistical procedure that demonstrates whether or not a comprehensive test battery measures the constructs it was designed to measure. The CFA results for the KABC-II are quite strong across the age range, showing that the revised test provides measurement of the theoretical constructs we intended to measure. Criterion-related validity indicates how well the KABC-II scales and subtests correlate with other measures of mental ability, such as the WISC-IV, and with tests of academic achievement (e. g., the Woodcock-Johnson III). The KABC-II is correlating substantially with these other measures — often in the 70s and 80s — supporting the validity of the KABC-II scales. The KABC-II is measuring the same key cognitive abilities as other major tests, and correlates substantially with school achievement. The question of fairness comes in when we examine how different ethnic groups perform relative to each other on a test. The K-ABC cut the difference between African-American children and Caucasian children in half, compared to traditional tests. We usually see Caucasian children scoring about 16 points higher than African-American children on traditional IQ tests. On the K-ABC, they scored about seven to eight points higher. So far, the data for the KABC-II are showing an even smaller difference. African-American children are performing relatively well on the KABC-II, as are Hispanic, American Indian, and Asian children. When we control the SES to the degree possible, the KABC-II yields substantially smaller differences than, for instance, the WISC-III.
The WISC-IV is now out, and while we don't yet have much data, we suspect that the difference between the WISC-IV and KABC-II will hold, with minority children scoring higher on the KABC-II. The only study we have right now is one Dr. Elaine Fletcher-Janzen conducted with a Taos Indian Pueblo population in New Mexico. She administered the KABC-II when the children in the study were almost eight years old. When the WISC-IV became available about eighteen months later, she administered it 30 children who were previously tested on the KABC-II. The scores between the two tests correlated substantially at about .85, so the tests are measuring the same major abilities. However, the children scored eight and one-half higher on the KABC-II, using the Luria model, than on the WISC-IV Full Scale IQ. Using the CHC model of the KABC-II, they scored seven and one-half points higher than on the WISC-IV Full Scale IQ.
Overall, the difference in scores we see on validity studies translates to Hispanic, American Indian, and Asian populations. The difference in the test's fairness is striking when compared to traditional tests.
Pearson: Will the KABC-II create as much opportunity for discussion and debate over the ways we think of and measure cognitive ability as its predecessor did?
Dr. Kaufman: Well, I'm not positive if generating controversy is good, but debate does occur when you have the courage to do something different.
When we published the original K-ABC, we separated core functions of ability, using a blend of Luria's neuropsychological model and Sperry's theories of psychobiology. This idea of a theory-based test was new at the time, and it provoked debate — even the inclusion of teaching items provoked debate, since no one had done it before. We took risks, went against the established order, and caused controversy.
People debated the reduced ethnic differences in our scores, as well. There was a lot of discussion over whether our scores indicated that we weren't measuring ability as well, or whether we were measuring ability differently.
Now, in 2004, many tests have incorporated aspects of the original K-ABC. Tests are not as rigid. Other tests are theory based, others include teaching items. And now we're doing things differently again. No one else has offered two global scores with one test. No one has incorporated two models and given the examiner a role in selecting the model that is best for the individual child.
I think we're moving the field forward. We believe we have a strong theory base, and we very strongly believe we have created a test that is geared toward the interests and needs of children. Helping children and adolescents is the goal of this test, so we've worked to be fair to all children who may be assessed with the KABC-II. We've included a nice balance of memory, learning, and problem-solving tasks. We've created more complex tasks, and we're still getting reduced cultural differences. Our critics, some of whom said the limited variety of tasks on the K-ABC was responsible for the reduced score gap between ethnic groups (for example, because there were too many tests of memory), will have to find a new reason for challenging the differences.
Pearson: What else would you like to say about test development and test usage?
Dr. Kaufman: This is an exciting time to come out with a new test for children and adolescents, partly because so much is going on nationally. Even as legislators are debating laws that would limit the use of cognitive ability, or IQ, testing to qualify children for services, we're providing an excellent measure of ability in five key areas. The KABC-II is useful in identifying strengths and weaknesses for learning disabled children. Regardless of laws, we believe that psychologists will need good tests for their work with children — particularly if legislation limits how ability testing may be used in the future for the assessment of children with possible learning disabilities.
We're also excited to be coming out with a good test when other good tests are available. In 1983, as the first theory-based test, we were unique. Now psychologists have choices. We feel that the KABC-II will be a strong choice that will help psychologists help children. And we feel wonderful about being able to help children.