Assessment of Cognitive Function A study with the Swedish version of the Kaufman Assessment Battery for Children

The Kaufman Assessment Battery for Children (K-ABC) is a test that measures cognitive fbnction in children. A Swedish version has been produced for use as a research instrument in a longitudinal follow-up study of children that needed neonatal intensive care at the Uppsala University Children's Hospital, Sweden. A study using this Swedish version was carried out in 26 healthy children 10 years of age attending a primary school in Uppsala, Sweden. The results showed that the Swedish version of the K-ABC well discriminates cognitive fbnction in children aged 10 years. The scores of all subtests were distributed over the scale and the mean scores mostly corresponded to near average or above average performance when compared with the mean values obtained in the standardization sample.


INTRODUCTION
The Kaufman Assessment Battery for Children (K-ABC) is a recently introduced individually administered test of intelligence and achievement, designed for use in assessments of children between the ages of 2.5 and 12.5 years (Kaufman, 1983a).
In the K-ABC, intelligence is defined in terms of a child's ability to process information in order to solve problems, and is based on the theory of neuropsychology and cognitive psychology.Intelligenceor problem solving abilityis measured with two Mental Processing Scales.One is the Sequential Processing Scale, which measures the child's capability of processing information serially or temporally in a stepwise fashion.The other one is the Simultaneous Processing Scale, in which the child is required to solve problems that are analogical, spatial and organizational in nature.A Mental Processing Composite Scale is a unification of the Sequential and Simultaneous Processing Scales and is intended for measuring total intelligence.The Achievement Scale in the K-ABC provides an estimate of previous learning by using subtests that measure verbal intelligence, general knowledge and acquired school skills.In addition, the K-ABC includes a Non-verbal Scale which serves to estimate intellectual potential in hearing-impaired or language-disordered children.
The theoretical basis of the test is related to recent advances in areas of cognitive development, learning theory and neuropsychology.These advances have been applied to various aspects of the test construction and validation, including modem stimulus materials (Kauhan, 1983b; Kamphaus & Reynolds, 1984).
The neuropsychological theory applied by Kau6nan & Kauhan (1993b) derives from two lines: the information processing approach of Alexander R. Luria, and the cerebral specialization theory.According to Luria's model the sequential-simultaneous dichotomy corresponds to the successive (fionto-temporal) and simultaneous (occipito-parietal) modes of information processing.According to the cerebral specialization theory, sequential processing corresponds to the ability of problem solving associated with the left hemisphere and simultaneous processing is believed to take place in the domain of the right hemisphere.In the majority of individuals, the left cerebral hemisphere appears to be specialized for linguistic, propositional, serial and analytical tasks and the right hemisphere for more non-verbal, synthetic and holistic tasks.Identification of significant strengths and weaknesses among the Mental Processing subtests and among the Achievement subtests offers, then, additional information for purposes of diagnosis and remediation (Kaufman et al, 1987;Reynolds et al, 1989).K-ABC is a test developed in the USA and standardized for use in American and German children (Kaufman, 1983a;Artner et af, 1989).It has also been adapted for British children, and at present it is the most commonly used cognitive test in international research on both normal and at-risk groups of children (Achenbah et al, 1993;Hack et al, 1994;abrzut et al, 1987;Roth et al, 1994;Ulissi et al, 1989).A total of 43 validity studies of the K-ABC have been described by Kaufman & Kaufman (1983b).Additionally, the K-ABC has been reported to give more detailed information about children's cognitive fknction as compared with other standardized tests, for instance the Wechsler Intelligence Scale for Children-Revised (Hack et al, 1994;Roth et al, 1993;Roth et al, 1994).
The purpose of the present study was to translate the K-ABC into Swedish and to evaluate the Swedish version of the test in a group of healthy children.It is intended that the Swedish version of the K-ABC shall be used as a research instrument in an epidemiological follow-up study of children who needed neonatal intensive care at the Uppsala University Children's Hospital, Sweden, and were born between 1 January 1986 and 30 April 1989 (Eizirik et al,   1994; Eizirik at al, 1996; Persson & Stromberg, 1995).The cognitive hnction of these children will be assessed in the spring of the year of their tenth birthday, when the majority of them will be attending the fourth year of the junior level of compulsory school.
Regardless of the parameters selected, standard scores have a constant meaning when based on distributions that are normal or approximately normal.Scores that are plus or minus one standard deviation from the mean include about 68 per cent of the individuals in a given population, a range of 2 standard deviations below and above the mean includes 95 per cent, and nearly everyone (over 99 per cent) is included between 3 standard deviations below and above the mean.Therefore 68 per cent or about two-thirds of normal children will earn standard scores of between 85 and 115 and scaled scores of between 7 and 13.Similarly, 95 per cent will score between 70 and 130 and virtually all the children will obtain scores in the 55 to 145 range; for the scaled scores the corresponding values are 4 to 16 and 1 to 19 (Kauhan & Kaufhan, 19b3b).

Translation of the K-ABC into Swedish and adaptation of items
The translation of the K-ABC into Swedish was done in collaboration with a school teacher with long experience in teaching in primary school.Special attention was paid to the Achievement Subtest Faces and Places, where 35 pictures of famous persons or places had to be changed to suit the Swedish context.Instead of presenting, for example, a picture of George Washington, we showed Gustav Vasa; instead of Hellen Keller, we showed Astrid Lindgren; instead of Ben Franklin, we showed Carl von Lim6 etc.

MATERIAL AND TEST PROCEDURE
A total of 26 children, 14 boys and 12 girls, participated in the study.The children were recruited from the junior level of compulsory school in Uppsala, Sweden, all of them attending the first term of the fourth year.Their ages varied between 9.06 and 10.06 years, with a mean of 9.08 years (median = 10.02 and mode =10.03 years).All children lived in the area close to the school.
The procedure for recruiting the children was as follows: First, contact with the school director and the teacher was established in order to explain the study procedures and to obtain their consent.Subsequently, the parents of all children in one class were informed by letter about the purposes of the assessment.As the participation in the study was voluntary, it was requested that the children should return the letter to the school teacher with the parents' decision regarding agreement to take part.Of a total of 29 children, 19 (65%) agreed to participate.Finally, 10 additional parents of children from another class, selected by a draw, were contacted and seven answered positively.No data on background characteristics of participants and non-participants are available.
All the assessments were performed in a quiet room at the school and on schedules defined in agreement with the children's parents.The assessment lasted about 90 minutes and the children were offered some juice and biscuits prior to the testing.The subtests Magic Window, Face Recognition and Expressive Vocabulary were not administered, as they are never included in assessments of children between 7.0 and 12.5 years of age (Kaufman & Kaufman, 1983a).

Global Scales
The scores of the Sequential Processing, Simultaneous Processing and Mental Processing Composite ranged between 2 standard deviations below and 2 above the mean of the American standardization sample, while the Achievement Scale showed scores within 1 standard deviation below and 3 standard deviations above the same mean (Table 1).The mean score of the Sequential Processing in the study corresponded to average performance and the mean scores of the remaining Global Scales to above average performance.

Simultaneous Processing Subtests
The scores of the Simultaneous Processing subtest Photo Series ranged between 2 standard deviations below and 2 above the mean of the American standardization sample (Table 3).
The scores of the Matrix Analogies subtest ranged between 2 standard deviations below and 3 above that mean, and the remaining subtests Gestalt Closure, Triangles and Spatial Memory showed scores between 1 standard deviation below and 2 standard deviations above the mean of the American standardization sample.The mean score for the subtests corresponded to a performance above average, except for Gestalt Closure and Spatial Memory, which were classified as average.

Achievement Scale
The scores for the Achievement subtests Arithmetic, Readingmecoding and Reading/ Understanding ranged between 1 standard deviation below and 3 standard deviations above the mean of the American standardization sample (Table 4).For the remaining subtests, Faces and Places and Riddles, the scores were within 2 standard deviations below and 2 above that mean and corresponded to average performance.The mean score of the Arithmetic subtest was above average; the mean score of Readinflnderstanding, well above average; and that of Reading/Decoding corresponded to upper extreme performance.
As the Achievement subtest ReadingiDecoding measures the child's ability to identify letters and to read and pronounce words, it is possible that the upper extreme performance in this subtest was related to a use of less complex Swedish words in the translation than those presented in the American version.A speech pathologist, with large experience in testing children, reviewed the Swedish words and made new suggestions, taking into account the numbers of syllables, word endings, common and uncommon words and also both phonetic and non-phonetic pronunciations, as suggested by Kaufman & Kaufman (1983b).The new words were tested in another 16 children who were attending the fourth class of the junior level, 13 of them from the same school as the children tested with the whole K-ABC.It was found that the mean score of the new version of the ReadingIDecoding subtest was lower than that obtained earlier (Table 5), but was well above the average performance in the US standardization sample.The scores varied between 1 standard deviation below and 3 standard deviations above the mean of the American standardization sample.The Swedish version of the K-ABC is intended for use as a research instrument in a longitudinal follow-up study of children who needed neonatal intensive care at the Uppsala university Children's Hospital, Sweden.The translation of the K-ABC into Swedish was focused not only on grammatical rules but also on changes of stimulus necessary to suit the Swedish context.
The results of the study showed that the scores of all K-ABC subtests were distributed over the scale, demonstrating that the Swedish version of the K-ABC discriminates cognitive fUnction in children at the age of 10 years.The mean scores of the K-ABC subtests were mostly close to average or above average performance as compared with the mean scores of the American standardization sample.
The tendency to higher performance among the Swedish children in comparison with American children might be due to differences in background characteristics of the sample.Although it was not the intention in the present study to collect demographic variables regarding the children, the sample might have differed, especially in socio-economic background, from the American national standardization sample (more than 2.000 children, tested at 34 test sites in 24 states).Parental education, for instance, might have been higher in the present study, as 19.2% of the parents from the standardization sample had less than highschool education, whereas in Sweden there is a nine-year compulsory school.Additionally, 27.5% of the American sample consisted of children from minority groups of Blacks, Hispanics, Native Americans, Asians, Alaskan Natives and Pacific Islanders, in contrast to our sample, which included mostly Swedish Caucasian children.Another aspect is that since many of the K-ABC subtests, especially those from the achievement scale, are a function of richness of early environment, cultural opportunities at home, availability of magazines and newspapers and school learning (Kaufman & Kaufman, 1983b), the supposed differences in socio-economic background between the two samples might have contributed to the observed differences in performance.In addition, it is difficult to say whether the sample from a central part of Uppsala, a university city, is fully representative for the performance of Swedish tenyear old children as a whole.
The K-ABC needs to undergo all the steps for standardization and validation for Swedish children as proposed by Kaufman & Kaufman (1983b) before it can become a generally available test in Sweden.Our impression, however, is that the K-ABC is an important innovation in the practice of cognitive assessment and a major contribution to neuropsychology and education.

CONCLUSION
The present study showed that the Swedish version of the K-ABC well discriminates cognitive function in a small selected sample of children at the age of 10 years.The scores of all subtests were distributed across the scale and the mean scores mostly corresponded to near average or above average performance when compared with the mean scores of the American standardization sample.The Swedish version of the K-ABC would thus seem to be a suitable research instrument for use in the planned follow-up study of children who needed neonatal intensive care and a group of healthy control children.

Table 1 . Standard Scores for the Global Scales (n=26)
The scaled scores for the Sequential Processing subtest Number Recall ranged within 2 standard deviations below and 2 above the mean of the American standardization sample (Table2).The scores for the remaining subtests, Hand Movement and Word Order, ranged between 1 standard deviation below and 2 standard deviations above the mean of the standardization sample.The mean score for all sequential subtests of the study lay at the level of average performance.

Table 5 . Standard Score for the Achievement Subtest ReadinglDecoding (n=16)
DISCUSSIONThe present study was undertaken to translate the K-ABC into Swedish and to test the Swedish version in a group of healthy children at the age of 10 years.The K-ABC is a recently introduced test (Kaufman & Kaufman, 1983b) which measures cognitive function in children.Its construction has a strong theoretical and research basis and it distinguishes acquired factual knowledge from the ability to solve unfamiliar problems.Additionally, the K-ABC includes novel tasks and has been reported to give more detailed information about children's cognitive function compared with other standardized tests(Hack etal, 1994; Roth et al, 1993; Roth et al, 1994).