Healthy Children Get Low Scores Too: Prevalence of Low Scores on the NEPSY-II in Preschoolers, Children, and Adolescents

Archives of Clinical Neuropsychology 25 (2010) 182 190 Healthy Children Get Low Scores Too: Prevalence of Low Scores on the NEPSY-II in Preschoolers, Children, and Adolescents Brian L. Brooks 1, *, Elisabeth M.S. Sherman 1, Grant L. Iverson 2 1 Alberta Children s Hospital, University of Calgary, Calgary, Alberta, Canada 2 Department of Psychiatry, British Columbia Mental Health and Addiction Services, University of British Columbia, Vancouver, British Columbia, Canada *Corresponding author at: Neurosciences Program, Alberta Children s Hospital, 2888 Shaganappi Trail NW, Calgary, AB, Canada T3B 6A8. Tel.: þ1-403-955-2597; fax: þ1-403-955-7045. E-mail address: brian.brooks@albertahealthservices.ca (B.L. Brooks). Accepted 24 January 2010 Abstract Knowing the prevalence of low test scores in healthy people is valuable for clinical interpretation of neuropsychological performance because it reduces the likelihood of over-diagnosing cognitive deficits. Base-rate information on adult batteries has flourished recently but is relatively unknown for pediatric tests. The purpose of this paper is to present the base rates of low scores for a pediatric neuropsychological battery, the NEPSY-II. Participants included 1,200 healthy preschoolers, children, and adolescents between 3 and 16 years of age from the NEPSY-II standardization sample. Measures included subtests from the attention and executive functioning, language, learning and memory, and visuospatial processing domains, organized to yield a 1- and 2-hr battery with optimal reliability. Analyses were conducted for three age groups (3 4, 5 6, and 7 16 years) and stratified by the level of parental education (11, 12, 13 15, and 16þ years). In 3 4-year-olds, it was uncommon to have three or more scores 10th percentile. For 5 6-year-olds, having three or more low scores on the 1- or 2-hr battery was uncommon. In 7 16-year-olds, it was uncommon to have four or more low scores on a 1-hr battery and uncommon to have five or more low scores on a 2-hr battery. With all age groups, the prevalence of low scores decreased substantially with higher levels of parental education. Consistent with the literature on base rates of low scores in adult batteries, having some low scores is common in healthy children. Look-up tables are provided to help in the clinical interpretation of low scores with preschoolers, children, and adolescents. Keywords: Neuropsychology; Base rates; Misdiagnosis; Pediatrics; Children; Abnormal scores Introduction Neuropsychologists administer several tests and interpret numerous test scores in any given assessment. As a result, large amounts of data are analyzed and multiple comparisons are undertaken. In statistical analyses, multiple comparisons of data require adjusted significance criteria to control the false-positive rate, which increases as more comparisons are made (i.e., the likelihood of making a Type I error, or to conclude that an effect is present when it occurs only by chance, increases as more variables are analyzed). However, in a neuropsychological assessment with numerous test scores, adjustments for multiple comparisons in order to control for false-positive rates are not employed. One method of adjustment in clinical practice, which lowers false-positive rates, involves interpreting scores with reference to the prevalence of low scores that occurs in healthy people when the entire battery is administered and interpreted. When interpreting numerous scores, knowing the expected base rates of low scores across the battery of tests is an important addition to clinical interpretation in order to control the false-positive rate and decreases the likelihood of misidentifying a commonly occurring low score as a clinically significant problem. The importance of understanding and utilizing base-rates data when interpreting performance on numerous test scores cannot be stressed enough. Clinicians are faced with interpreting a large number of scores and must determine if the results reflect cognitive impairment, whether acquired or arising from a developmental condition. Do some low scores necessarily # The Author 2010. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org. doi:10.1093/arclin/acq005 Advance Access publication on 22 February 2010

indicate acquired cognitive impairment secondary to a damaged or dysfunctional central nervous system or may some of these low scores be expected in healthy children? If so, how many low scores would be considered common and how many would be considered uncommon? Currently, there exist little to no published data to help answer these questions. The base rates of low scores have been studied in several adult neuropsychological batteries. These include the Halstead Reitan Neuropsychological Battery (Binder, Iverson, & Brooks, 2009; Heaton, Grant, & Matthews, 1991; Heaton, Miller, Taylor, & Grant, 2004), the Neuropsychological Assessment Battery (Brooks, Iverson, & White, 2009; Iverson & Brooks, in press; Iverson, Brooks, White, & Stern, 2008), the Wechsler Adult Intelligence Scale Third Edition/Wechsler Memory Scales Third Edition (Iverson, Brooks, & Holdnack, 2009), and different flexible batteries (Palmer, Boone, Lesser, & Wohl, 1998; Schretlen, Testa, Winicki, Pearlson, & Gordon, 2008). The prevalence of low scores has also been reported for tests within specific cognitive domains, such as memory performance on the WMS-III (Brooks, Iverson, Holdnack, & Feldman, 2008) and the NAB Memory Module (Brooks, Iverson, & White, 2007). Although there has been accumulating research and publication of clinically useful tables for multiple test interpretation in adults, relatively little has been studied for pediatric neuropsychology. Brooks, Iverson, Sherman, and Holdnack (2009) presented the prevalence of low scores on the Children s Memory Scale (CMS; Cohen, 1997). Similar to adult memory batteries, Brooks and colleagues found that having a low score is relatively common in healthy children and adolescents and that having low scores becomes less common in those children and adolescents with higher intelligence. The NEPSY-II is comprised of several neuropsychological subtests, appropriate for children between 3 and 16 years of age, which cover cognitive domains that are of key interest as part of a neuropsychological assessment within a pediatric neurology clinic. The purpose of this research is, therefore, to examine and present the base rates of low scores on the NEPSY-II in healthy preschoolers, children, and adolescents (Korkman, Kirk, & Kemp, 2007a). As part of this research, clinically useful tables will be presented that can facilitate the interpretation of test scores (and maintain a desired false-positive rate) when a battery of NEPSY-II tests is administered. Materials and Methods Participants The NEPSY-II normative sample is a national, stratified, random sample consisting of 1,200 preschoolers, children, and adolescents between the ages of 3 and 16 years. Standardization of the NEPSY-II occurred between 2005 and 2006. There are 100 children (50 boys and 50 girls) in each of 12 age groups: 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13 14, and 15 16 years of age. For ages 3 12 years, each age group contains 50 children in the first 6 months and 50 children in the second 6 months of the year. For adolescents between 13 and 16 years, there are 50 for each year of age. Stratification of the entire normative sample by age, race/ethnicity, geographic location, and parental education was based on the October 2003 U.S. Census Survey. As illustrated in Table 3.1 of the Clinical and Interpretive Manual (Korkman, Kirk, & Kemp, 2007c), each age group was sampled to closely resemble the 2003 U.S. Census information. Exclusion criteria included diagnosis of a number of conditions that could potentially affect scores (i.e., neurological, learning, sensory/motor, or psychiatric disorders), as well as English as a second language, recent history of previous testing, and medication usage that might potentially impact performance. Measures B.L. Brooks et al. / Archives of Clinical Neuropsychology 25 (2010) 182 190 183 The NEPSY-II (Korkman et al., 2007a) is a comprehensive, co-normed, and multidomain neuropsychological battery designed for assessing neurocognitive abilities in preschoolers, children, and adolescents. It is a flexible battery of 32 subtests that permits the administration of specific subtests, groups of subtests, or the entire battery. The NEPSY-II is divided into six theoretically derived domains of cognitive functioning: Attention and Executive Functioning; Language; Memory and Learning; Sensorimotor; Social Perception; and Visuospatial Processing. From our experience, few clinicians would consider administering the entire NEPSY-II in an assessment. Most often, selected NEPSY-II subtests are included as part of a larger neuropsychological assessment. For the purpose of this study, only a selection of NEPSY-II subtests was considered for the analyses. These subtests, which were chosen according to clinical and psychometric rationale (see Brooks, Sherman, & Strauss, 2010), cover a broad range of cognitive abilities, including attention and executive functioning, language, learning and memory, and visuospatial abilities (see Table 1 for a list of the subtests in 3 4-, 5 6-, and 7 16-year-olds). Selection of some subtests was limited for some age groups because some tests are not applicable and not normed for younger children (e.g., Inhibition Switching is normed for 7 16-year-olds only and cannot be administered to younger children). Despite these restrictions, consistency was attempted across the batteries

184 B.L. Brooks et al. / Archives of Clinical Neuropsychology 25 (2010) 182 190 Table 1. NEPSY-II subtests for a 1- or 2-hr battery with children and adolescents NEPSY-II domains and subtests 3 4-year-olds 5 6-year-olds 7 16-year-olds 1-hr 2-hr 1-hr 2-hr Attention and Executive Functioning Animal Sorting Total Correct Sorts 3 Auditory Attention Total Correct 3 3 Response Set Total Correct 3 Inhibition: Naming Total Completion Time 3 3 3 3 Inhibition: Inhibition Total Completion Time 3 3 3 3 Inhibition: Switching Total Completion Time 3 3 Language Comprehension of Instructions Total 3 3 3 3 3 Phonological Processing Total 3 3 3 Speeded Naming Total Completion Time 3 3 3 Memory and Learning Memory for Designs Total 3 3 3 3 3 Memory for Designs Delayed Total 3 3 3 3 Narrative Memory Free & Cued Recall Total 3 3 3 3 3 Narrative Memory Free Recall Total 3 3 3 3 Word List Interference Repetition Total 3 3 Word List Interference Recall Total 3 3 Visuospatial Processing Block Construction Total Score 3 3 3 3 3 Geometric Puzzles Total Score 3 3 3 Notes: NEPSY-II subtests were chosen based on overlap across age groups, coverage of cognitive domains, clinical utility, and reliability evidence summarized in Korkman and colleagues (2007b) and Brooks, Strauss, and colleagues (2009). Times for the NEPSY-II batteries are only estimates, based on times provided in Korkman and colleagues (2007b) and derived from the normative sample. Times with clinical populations could be different. by (a) including many of the same tests across age groups, and (b) including all of the tests from shorter batteries within the longer batteries. For 3 4-year-olds, the battery of tests covers language, learning and memory, and visuospatial abilities and takes approximately 1 hr to administer (administration times were derived from Table 2.2 in the NEPSY-II Administration Manual; Korkman, Kirk, & Kemp, 2007b). These subtests yield 7 age-adjusted scores. The subtests included for the 3 4-year-olds are also included in the longer batteries for the older age groups. For the 5 6- and 7 16-year-olds, two batteries were considered. The shorter batteries for these two age groups take approximately 1 hr to administer and include subtests from the attention and executive functioning, language, learning and memory, and visuospatial domains. In the 5 6-year-olds, there are 8 scores derived from the 1-hr battery. In the 7 16-year-olds, there are 11 scores derived from the 1-hr battery. The longer batteries for these two age groups, which includes the same subtests from the 1-hr batteries and adds several additional subtests to round out the coverage of the cognitive domains, take approximately 1.5 2 hr to administer. In the 5 6-year-olds, there are 12 scores derived from the 2-hr battery. In the 7 16-year-olds, there are 17 scores derived from the 2-hr battery. Although administrations times may be longer in some clinical groups, for the sake of simplicity, these test groupings will be referred to as either 1-hr or 2-hr batteries, respectively. For the NEPSY-II normative sample, parental education was obtained by having parents fill out a questionnaire that included the highest level of formal education completed by each parent living in the household (e.g., less than high school is 11 years; high school diploma ¼ 12 years; college diploma ¼ 14 years; bachelor s degree ¼ 16 years; master s degree or other post-bachelor s degree education ¼ 16þ years). For children in a two-parent household, parental education was recorded as the average of the two levels of education (see p. 41 in Korkman, et al., 2007b, for additional information). Analyses The base rates of low scores for the three age groups were calculated by considering performance on all subtests, within the specified battery, simultaneously. A number of cutoff scores were considered to allow flexibility in the use of base-rate tables for different clinical purposes. Cutoff scores included: 25th percentile (i.e., scaled score 8 or percentile rank 11th 25th); 10th percentile (i.e., scaled score 6 or percentile rank 6th 10th); 5th percentile (i.e., scaled score 5 or percentile rank 2nd 5th); and,2nd percentile (i.e., scaled score 3 or percentile rank,2nd).

B.L. Brooks et al. / Archives of Clinical Neuropsychology 25 (2010) 182 190 185 Descriptive analyses of the prevalence of low NEPSY-II subtest scores are presented for each of the three age groups (i.e., 3 4-, 5 6-, and 7 16-year-olds), with results presented for the total age-group sample as well as stratified by the level of parental education (i.e., 11, 12, 13 15, and 16þ years). Results The base rates of low NEPSY-II subtest scores in 3 4-year-olds are presented in Table 2. When the seven scores are considered simultaneously, 71.5% of healthy preschoolers had one or more scores 25th percentile, 40.5% had one or more scores 10th percentile, 24.7% had one or more scores 5th percentile, and 8.9% had one or more scores below the 2nd percentile. Considering the base rates in Table 2 another way, it was uncommon (i.e., less than or equal to 10% of the sample) for 3 4-year-olds to have five or more low scores 25th percentile, three or more scores 10th percentile, two or more scores 5th percentile, and one or more scores,2nd percentile. The base rates of low NEPSY-II scores in preschool children decrease as the level of parental education increases. For example, in those preschoolers with parental education of 11 years or less, 65.0% had 1 or more scores 10th percentile compared with 29.8% of those preschoolers who have parents with 16 or more years of education [x 2 (1) ¼ 7.23, p ¼.007, odds ratio ¼ 4.44 (95% CI ¼ 1.5 13.0)]. The base rates of low NEPSY-II subtest scores in 5 6-year-olds, considered for two different battery lengths, are presented in Table 3. For the 1-hr battery, when the eight scores are considered simultaneously, 70.3% of healthy children had one or more scores 25th percentile, 37.2% had one or more scores 10th percentile, 20.7% had one or more scores 5th percentile, and 4.8% had one or more scores below the 2nd percentile. On the 1-hr battery, it was uncommon for 5 6-year-olds to have six Table 2. Base rates of low NEPSY-II scores in 3 4-year-old children Number of low NEPSY-II scores Total sample Parent education (years) 11 12 13 15 16þ 25th percentile Six or more 1.3 10.0 Five or more 6.3 15.0 7.0 4.2 4.3 Four or more 17.1 35.0 20.9 16.7 6.4 Three or more 29.7 45.0 30.2 29.2 23.4 Two or more 51.3 70.0 58.1 50.0 38.3 One or more 71.5 85.0 76.7 72.9 59.6 No low scores 28.5 15.0 23.3 27.1 40.4 10th percentile Four or more 1.9 5.0 4.7 Three or more 6.3 20.0 7.0 4.2 2.1 Two or more 16.5 30.0 16.3 18.8 8.5 One or more 40.5 65.0 44.2 37.5 29.8 No low scores 59.5 35.0 55.8 62.5 70.2 5th percentile Four or more 0.6 2.3 Three or more 0.6 2.3 Two or more 6.3 10.0 9.3 6.3 2.1 One or more 24.7 55.0 30.2 18.8 12.8 No low scores 75.3 45.0 69.8 81.3 87.2,2nd percentile Three or more 0.6 2.3 Two or more 3.2 5.0 7.0 2.1 One or more 8.9 5.0 16.3 8.3 4.3 No low scores 91.1 95.0 83.7 91.7 95.7 Notes: Cumulative percentages presented in bold represent an uncommon number of low scores, based on prevalence rates of 10% of the sample (i.e., a 10% false-positive rate). Values represent cumulative percentage of each sample. These are slight variations due to rounding of numbers. In 3 4-year-old children, there are seven scores considered for these analyses. Some NEPSY-II scores have skewed distributions in some age groups and are presented as percentile ranks. Percentile ranks are divided into the following categories:,2, 2 5, 6 10, 11 25, and 25 50, 51 75, and 76þ. Cutoff scores included: 25th percentile, SS 8, and percentile rank ¼ 11 25; 10th percentile, SS 6, and percentile rank ¼ 6 10; 5th percentile, SS 5, and percentile rank ¼ 2 5;,2nd percentile, SS, 4, and percentile rank, 2. Sample sizes included: Total sample, n ¼ 158; 11 years, n ¼ 20; 12 years, n ¼ 43; 13 15 years, n ¼ 48; and 16þ years, n ¼ 47. For children in a two-parent household, parental education was recorded as the average of the two levels of education. Standardization data from the NEPSY, Second Edition (NEPSY-II).

186 B.L. Brooks et al. / Archives of Clinical Neuropsychology 25 (2010) 182 190 Table 3. Base rates of low NEPSY-II scores in 5 6-year-old children Number of low NEPSY-II scores 1-hr battery 2-hr battery Total sample Parent education (years) Total sample Parent education (years) 11 12 13 15 16þ 11 12 13 15 16þ 25th percentile Ten or more 2.2 3.2 4.3 Nine or more 5.1 6.7 9.7 6.4 Eight or more 0.7 5.9 5.1 6.7 9.7 6.4 Seven or more 3.4 5.9 9.1 2.0 9.4 13.3 12.9 14.9 Six or more 6.2 11.8 12.1 6.0 14.5 13.3 16.1 23.4 4.4 Five or more 14.5 17.6 24.2 18.0 2.2 26.1 40.0 35.5 29.8 11.1 Four or more 26.2 41.2 42.4 28.0 6.7 34.8 53.3 48.4 36.2 17.8 Three or more 38.6 58.8 54.5 36.0 22.2 46.4 66.7 54.8 48.9 31.1 Two or more 55.2 88.2 63.6 54.0 37.8 61.6 93.3 71.0 59.6 46.7 One or more 70.3 88.2 81.8 70.0 55.6 82.6 100 93.5 83.0 68.9 No low scores 29.7 11.8 18.2 30.0 44.4 17.4 0.0 6.5 17.0 31.1 10th percentile Six or more 1.4 3.2 2.1 Five or more 0.7 3.0 2.2 3.2 4.3 Four or more 2.8 5.9 3.0 4.0 5.8 6.7 6.5 10.6 Three or more 6.9 17.6 6.1 10.0 10.1 6.7 9.7 21.3 Two or more 17.2 29.4 21.2 24.0 2.2 20.3 33.3 16.1 27.7 11.1 One or more 37.2 58.8 36.4 40.0 26.7 49.3 66.7 48.4 48.9 44.4 No low scores 62.8 41.2 63.6 60.0 73.3 50.7 33.3 51.6 51.1 55.6 5th percentile Four or more 0.7 5.9 2.2 6.7 3.2 2.1 Three or more 2.1 5.9 3.0 2.0 4.3 6.7 3.2 8.5 Two or more 9.0 29.4 9.1 10.0 8.7 20.0 9.7 12.8 One or more 20.7 35.3 27.3 24.0 6.7 29.7 40.0 32.3 34.0 20.0 No low scores 79.3 64.7 72.7 76.0 93.3 70.3 60.0 67.7 66.0 80.0,2nd percentile Two or more 1.4 5.9 2.0 2.2 6.7 4.3 One or more 4.8 11.8 6.1 6.0 10.1 13.3 9.7 12.8 6.7 No low scores 95.2 88.2 93.9 94.0 100 89.9 86.7 90.3 87.2 93.3 Notes: Cumulative percentages presented in bold represent an uncommon number of low scores, based on prevalence rates of 10% of the sample (i.e., a 10% false-positive rate). Values represent cumulative percentage of each sample. These are slight variations due to rounding of numbers. In 5 6-year-old children, there are 8 scores for the 1-hr battery and 12 scores for the 2-hr battery that were considered for these analyses. Some NEPSY-II scores have skewed distributions in some age groups and are presented as percentile ranks. Percentile ranks are divided into the following categories:,2, 2 5, 6 10, 11 25, and 25 50, 51 75, and 76þ. Cutoff scores included: 25th percentile, SS 8, and percentile rank ¼ 11 25; 10th percentile, SS 6, and percentile rank ¼ 6 10; 5th percentile, SS 5, and percentile rank ¼ 2 5;,2nd percentile, SS, 4, and percentile rank, 2. Sample sizes for the 1-hr battery included: Total sample, n ¼ 145; 11 years, n ¼ 17; 12 years, n ¼ 33; 13 15 years, n ¼ 50; and 16þ years, n ¼ 45. Sample sizes for the 2-hr battery included: Total sample, n ¼ 138; 11 years, n ¼ 15; 12 years, n ¼ 31; 13 15 years, n ¼ 47; and 16þ years, n ¼ 45. For children in a two-parent household, parental education was recorded as the average of the two levels of education. Standardization data from the NEPSY, Second Edition (NEPSY-II). or more low scores 25th percentile, three or more scores 10th percentile, two or more scores 5th percentile, and one or more scores,2nd percentile. On the 2-hr NEPSY-II battery (12 scores), 82.6% of healthy children had one or more scores 25th percentile, 49.3% had one or more scores 10th percentile, 29.7% had one or more scores 5th percentile, and 10.1% had one or more scores below the 2nd percentile. On the 2-hr battery, it was uncommon for 5 6-year-olds to have seven or more low scores 25th percentile, three or more scores 10th percentile, two or more scores 5th percentile, and one or more scores,2nd percentile. As was the case for the younger group, the base rates of low NEPSY-II scores in 5 6-year-olds decreases as the level of parental education increases. In those children with parental education of 11 years or less, 58.8% had one or more scores 10th percentile on the 1-hr and 66.7% had one or more scores 10th percentile on the 2-hr battery. In comparison to those children with parents that have 16þ years of education, 26.7% had one or more low scores on the 1-hr battery [x 2 (1) ¼ 5.57, p ¼.018, odds ratio ¼ 3.9 (95% CI ¼ 1.2 12.4)]. Although only 44.4% of those with higher educated parents had one or more scores 10th percentile on the 2-hr battery (compared with 66.7%), this difference was not statistically significant [x 2 (1) ¼ 2.22, p ¼.14, odds ratio ¼ 2.5 (95% CI ¼ 0.8 8.2)].

B.L. Brooks et al. / Archives of Clinical Neuropsychology 25 (2010) 182 190 187 Table 4. Base rates of low NEPSY-II scores in 7 16-year-old children and adolescents Number of low NEPSY-II scores 1-hr battery 2-hr battery Total sample Parent education (years) Total sample Parent education (years) 11 12 13 15 16þ 11 12 13 15 16þ 25th percentile Fifteen or more 0.2 0.5 Fourteen or more 0.5 1.4 0.5 Thirteen or more 1.1 1.6 2.1 1.1 Twelve or more 1.6 4.7 2.1 1.6 Eleven or more 0.2 1.4 4.9 12.5 4.9 4.9 1.8 Ten or more 0.5 1.4 0.7 0.5 8.3 25.0 8.3 6.0 4.2 Nine or more 1.7 5.6 2.0 1.0 0.6 11.3 35.9 10.4 8.2 6.1 Eight or more 4.4 9.7 5.3 3.5 2.3 18.2 46.9 19.4 15.3 9.1 Seven or more 8.2 16.7 10.5 6.0 5.3 25.2 53.1 27.8 24.6 12.7 Six or more 16.3 31.9 17.8 14.6 10.5 34.4 68.8 40.3 31.1 19.4 Five or more 25.8 52.8 30.3 24.1 12.3 44.6 70.3 51.4 44.8 28.5 Four or more 38.7 63.9 46.1 37.2 23.4 56.8 84.4 63.9 54.1 43.0 Three or more 54.5 80.6 61.8 55.8 35.7 70.5 90.6 76.4 71.6 56.4 Two or more 69.7 90.3 77.0 69.3 55.0 79.3 93.8 87.5 78.1 67.9 One or more 84.7 94.4 88.8 85.9 75.4 92.1 96.9 91.0 93.4 89.7 No low scores 15.3 5.6 11.2 14.1 24.6 7.9 3.1 9.0 6.6 10.3 10th percentile Ten or more 0.2 1.6 Nine or more 0.2 1.4 0.4 1.6 0.5 Eight or more 0.2 1.4 0.9 1.6 0.7 1.1 0.6 Seven or more 0.3 1.4 0.6 2.9 6.3 2.8 2.7 1.8 Six or more 0.7 1.4 0.5 1.2 4.1 10.9 4.2 3.8 1.8 Five or more 2.4 4.2 1.3 3.0 1.8 8.6 18.8 9.7 9.8 2.4 Four or more 7.7 15.3 7.2 9.0 3.5 15.1 34.4 18.8 13.7 6.1 Three or more 16.0 29.2 17.8 15.6 9.4 25.4 46.9 29.9 24.0 14.5 Two or more 30.1 50.0 36.2 28.6 18.1 40.6 71.9 44.4 39.3 26.7 One or more 52.4 75.0 58.6 50.8 39.2 62.9 84.4 64.6 61.7 54.5 No low scores 47.6 25.0 41.4 49.2 60.8 37.1 15.6 35.4 38.3 45.5 5th percentile Eight or more 0.2 1.6 Seven or more 0.2 1.4 0.2 1.6 Six or more 0.2 1.4 0.7 4.7 0.5 Five or more 0.7 2.8 1.0 2.2 7.8 1.4 2.2 0.6 Four or more 2.0 6.9 2.0 2.0 5.2 15.6 6.9 3.8 1.2 Three or more 5.2 12.5 5.3 4.0 3.5 10.4 25.0 13.2 8.2 4.8 Two or more 14.8 23.6 21.1 11.6 9.4 24.1 45.3 27.8 21.9 15.2 One or more 34.8 52.8 38.2 33.7 25.7 44.1 67.2 46.5 43.2 33.9 No low scores 65.2 47.2 61.8 66.3 74.3 55.9 32.8 53.5 56.8 66.1,2nd percentile Three or more 0.4 0.7 0.5 Two or more 1.7 6.9 0.7 1.5 0.6 2.5 7.8 1.4 2.2 1.8 One or more 10.3 15.3 9.2 10.6 8.8 14.7 31.3 13.9 14.8 9.1 No low scores 89.7 84.7 90.8 89.4 91.2 85.3 68.8 86.1 85.2 90.9 Notes: Cumulative percentages presented in bold represent an uncommon number of low scores, based on prevalence rates of 10% of the sample (i.e., a 10% false-positive rate). Values represent cumulative percentage of each sample. These are slight variations due to rounding of numbers. In 7 16-year-old children, there are 11 scores for the 1-hr battery and 17 scores for the 2-hr battery that were considered for these analyses. Some NEPSY-II scores have skewed distributions in some age groups and are presented as percentile ranks. Percentile ranks are divided into the following categories:,2, 2 5, 6 10, 11 25, and 25 50, 51 75, and 76þ. Cutoff scores included: 25th percentile, SS 8, and percentile rank ¼ 11 25; 10th percentile, SS 6, and percentile rank ¼ 6 10; 5th percentile, SS 5, and percentile rank ¼ 2 5;,2nd percentile, SS, 4, and percentile rank, 2. Sample sizes for the 1-hr battery included: Total sample, n ¼ 594; 11 years, n ¼ 72; 12 years, n ¼ 152; 13 15 years, n ¼ 199; and 16þ years, n ¼ 171. Sample sizes for the 2-hr battery included: Total sample, n ¼ 556; 11 years, n ¼ 64; 12 years, n ¼ 144; 13 15 years, n ¼ 183; and 16þ years, n ¼ 165. For children in a two-parent household, parental education was recorded as the average of the two levels of education. Standardization data from the NEPSY, Second Edition (NEPSY-II).

188 B.L. Brooks et al. / Archives of Clinical Neuropsychology 25 (2010) 182 190 The base rates of low NEPSY-II subtest scores in 7 16-year-olds, considered for two different battery lengths, are presented in Table 4. For the 1-hr battery, when the 11 scores are considered simultaneously, 84.7% of healthy children and adolescents had one or more scores 25th percentile, 52.4% had one or more scores 10th percentile, 34.8% had one or more scores 5th percentile, and 10.3% had one or more scores below the 2nd percentile. On the 1-hr battery, it was uncommon for 7 16-year-olds to have seven or more low scores 25th percentile, four or more scores 10th percentile, three or more scores 5th percentile, and one or more scores,2nd percentile. On the 2-hr NEPSY-II battery for 7 16-year-olds (17 scores), 92.1% of healthy children and adolescents had one or more scores 25th percentile, 62.9% had one or more scores 10th percentile, 44.1% had one or more scores 5th percentile, and 14.7% had one or more scores below the 2nd percentile. On the 2-hr battery, it was uncommon for 7 16-year-olds to have 10 or more low scores 25th percentile, 5 or more scores 10th percentile, 3 or more scores 5th percentile, and 2 or more scores,2nd percentile. Consistent with the other age groups, the base rates of low NEPSY-II scores in 7 16-year-olds decreases with increasing parental education. Having one or more scores 10th percentile on the 1-hr battery is found in 75.0% of children and adolescents with parental education of 11 years or less compared with 39.2% of children and adolescents with parents that have 16þ years of education [x 2 (1) ¼ 26.00, p,.001, odds ratio ¼ 4.7 (95% CI ¼ 2.5 8.6)]. On the 2-hr battery, 84.4% had one or more scores 10th percentile compared with 54.5% of children and adolescents with parents who have 16þ years of education [x 2 (1) ¼ 17.58, p,.001, odds ratio ¼ 4.5 (95% CI ¼ 2.2 9.3)]. Discussion Few clinicians would argue that a single, isolated low score is solely indicative of cognitive impairment. This is partially because it is well known that obtaining a low score, particularly across a battery of neuropsychological tests (whether a fixed or flexible battery), is common in healthy people without impairments (e.g., Axelrod & Wall, 2007; Binder et al., 2009; Brooks et al., 2007, 2008; Brooks, Iverson, & White, 2009; Crawford, Garthwaite, & Gault, 2007; Heaton et al., 1991, 2004; Iverson & Brooks, in press; Iverson, Brooks, & Holdnack, 2008; Iverson, Brooks, White, et al., 2008; Palmer et al., 1998; Schretlen et al., 2008). However, is it common or uncommon to obtain two, three, or even four low scores? What if the scores fall below more stringent cutoffs for cognitive impairment (e.g., 2nd percentile)? What if the low score is obtained by a person who has high premorbid intelligence? Unfortunately, clinical judgment alone cannot adequately answer these types of questions because the number of test scores below any psychometric cutoff changes when multiple tests are administered and interpreted. Using the prevalence of low scores in healthy children provides an interpretive benchmark for common or uncommon test performance, akin to an adjustment for multiple comparisons within overall clinical interpretation. The results of this study further support the psychometric principles of multiple score interpretation discussed in Brooks, Strauss, Sherman, Iverson, and Slick (2009) and Iverson & Brooks (in press). First, low scores are common in healthy people. Second, the number of low scores depends on the cutoff score utilized. Third, the number of low scores depends on the number of tests administered and interpreted. Fourth, the number of low scores varies by a person s level of (presumed) functioning. In the present study, NEPSY-II scores were stratified by the level of parental education, a proxy estimate of expected cognitive level. This is the first study to examine children s neuropsychological test performance stratified by parental education. There is some literature that supports increased parental education being related to higher levels of intelligence in healthy children (e.g., Schoenberg, Lange, & Saklofske, 2007; van der Sluis, Willemsen, de Geus, Boomsma, & Posthuma, 2008) and children of mothers with epilepsy (e.g., Thomas, Sukumaran, Lukose, George, & Sarma, 2007). There are some advantages to stratification by parent education compared with stratification by other intellectual or cognitive scores. In particular, if a child has experienced a developmental or acquired neurological impairment that results in diminished intelligence, then using the (potentially lower) obtained intelligence score would increase the chance of saying that the number of low scores is broadly normal (i.e., making a Type II, or false negative, error). The positive relationship between parental education and child neurocognitive functioning supports the use of parental education-stratified tables in clinical practice, although caution should be exercised because parental intelligence, which can be a proxy of educational attainment, only accounts for about 17% of the variance in child s intelligence (Devlin, Daniels, & Roeder, 1997). Our recommendation is to consider both the unadjusted (total sample) and adjusted (parental-education stratified) base-rate information for making conclusions about a child s functioning compared with healthy children, because these address slightly different clinical questions. There are some limitations with this research that should be considered. First, the base rates of low scores presented are applicable only if the exact same subtests are administered and interpreted. In other words, proper use of the base-rate tables means that a different NEPSY-II subtest cannot be substituted in place of one of the subtests presented in Table 1 and a subtest cannot be omitted from the interpretation. It is hopeful that test publishers will eventually provide the base rates of low scores, for any combination of scores, as part of the sophisticated scoring programs that are included with tests. Second, the base rates of low scores in this study apply to the NEPSY-II, but it is well known that several other tests

are also administered and interpreted as part of a neuropsychological assessment. As such, the number of scores interpreted is typically greater, which likely increases the overall prevalence of low scores beyond the numbers presented in Tables 2 4. Third, the base rates are not stratified by the level of intelligence because the sample sizes of the concurrent validity subsamples who received all of the NEPSY-II subtests in this study were too small. Although the stratification by parental education has some unique advantages, having stratification by a child s intelligence would have also provided useful information. The overall goal of this study is to improve the diagnostic accuracy of neuropsychological measures through psychometrically based methods of interpretation. It is our intention to facilitate the clinical interpretation of test performance and to minimize the likelihood of misdiagnosing cognitive impairment when the number of low scores is actually common in healthy preschoolers, children, or adolescents. Future researchers should examine if there are differences in the base rates of low scores across various clinical samples (e.g., Attention-Deficit/Hyperactivity Disorder, learning disability, traumatic brain injury) that could be used to distinguish between groups. Conflict of Interest The authors have no known, perceived, or actual conflict of interest with this research. Data were provided to the first author by NCS Pearson, Inc. in order to conduct research, but this study was unfunded. Acknowledgements Standardization data from the NEPSY, Second Edition (NEPSY-II). Copyright # 2007 NCS Pearson, Inc. Used with permission. All rights reserved. The authors thank Helen Carlson, PhD, and Jennifer Bernardo for assistance with manuscript preparation. References B.L. Brooks et al. / Archives of Clinical Neuropsychology 25 (2010) 182 190 189 Axelrod, B. N., & Wall, J. R. (2007). Expectancy of impaired neuropsychological test scores in a non-clinical sample. International Journal of Neuroscience, 117 (11), 1591 1602. Binder, L. M., Iverson, G. L., & Brooks, B. L. (2009). To err is human: Abnormal neuropsychological scores and variability are common in healthy adults. Archives of Clinical Neuropsychology, 24, 31 46. Brooks, B. L., Iverson, G. L., Holdnack, J. A., & Feldman, H. H. (2008). Potential for misclassification of mild cognitive impairment: A study of memory scores on the Wechsler Memory Scale-III in healthy older adults. Journal of the International Neuropsychological Society, 14 (3), 463 478. Brooks, B. L., Iverson, G. L., Sherman, E. M. S., & Holdnack, J. A. (2009). Healthy children and adolescents obtain some low scores across a battery of memory tests. Journal of the International Neuropsychological Society, 15 (4), 613 617. Brooks, B. L., Iverson, G. L., & White, T. (2007). Substantial risk of Accidental MCI in healthy older adults: Base rates of low memory scores in neuropsychological assessment. Journal of the International Neuropsychological Society, 13 (3), 490 500. Brooks, B. L., Iverson, G. L., & White, T. (2009). Advanced interpretation of the neuropsychological assessment battery (NAB) with older adults: Base rate analyses, discrepancy scores, and interpreting change. Archives of Clinical Neuropsychology, 24 (7), 647 657. Brooks, B. L., Sherman, E. M., & Strauss, E. (2010). NEPSY-II: A developmental neuropsychological assessment, second edition. Child Neuropsychology, 16, 80 81. Brooks, B. L., Strauss, E., Sherman, E. M. S., Iverson, G. L., & Slick, D. J. (2009). Developments in neuropsychological assessment: Refining psychometric and clinical interpretive methods. Canadian Psychology, 50 (3), 196 209. Cohen, M. J. (1997). Children s Memory Scale. San Antonio, TX: The Psychological Corporation. Crawford, J. R., Garthwaite, P. H., & Gault, C. B. (2007). Estimating the percentage of the population with abnormally low scores (or abnormally large score differences) on standardized neuropsychological test batteries: a generic method with applications. Neuropsychology, 21 (4), 419 430. Devlin, B., Daniels, M., & Roeder, K. (1997). The heritability of IQ. Nature, 388, 468 471. Heaton, R. K., Grant, I., & Matthews, C. G. (1991). Comprehensive norms for an extended Halstead Reitan Battery: Demographic corrections, research findings, and clinical applications. Odessa, FL: Psychological Assessment Resources, Inc. Heaton, R. K., Miller, S. W., Taylor, M. J., & Grant, I. (2004). Revised comprehensive norms for an expanded Halstead Reitan Battery: Demographically adjusted neuropsychological norms for African American and Caucasian adults professional manual. Lutz, FL: Psychological Assessment Resources, Inc. Iverson, G. L., & Brooks, B. L. (in press). Improving accuracy for identifying cognitive impairment. In M. R. Schoenberg, & J. G. Scott (Eds.), The black book of neuropsychology: A syndrome-based approach. Iverson, G. L., Brooks, B. L., & Holdnack, J. A. (2008). Misdiagnosis of cognitive impairment in forensic neuropsychology. In R. L. Heilbronner (Ed.), Neuropsychology in the courtroom: Expert analysis of reports and testimony (pp. 243 266). New York: Guilford Press. Iverson, G. L., Brooks, B. L., & Holdnack, J. A. (2009). Low scores across the WAIS-III/WMS-III vary by education & intelligence, Paper presented at the 37th Annual Meeting of the International Neuropsychological Society, Atlanta, GA. Iverson, G. L., Brooks, B. L., White, T., & Stern, R. A. (2008). Neuropsychological assessment battery (NAB): Introduction and advanced interpretation. In A. M. Horton Jr, & D. Wedding (Eds.), The Neuropsychology Handbook (3rd ed., pp. 279 343). New York: Springer Publishing Inc. Korkman, M., Kirk, U., & Kemp, S. (2007a). NEPSY-II: A developmental neuropsychological assessment. San Antonio, TX: The Psychological Corporation. Korkman, M., Kirk, U., & Kemp, S. (2007b). NEPSY-II: Administration manual. San Antonio, TX: The Psychological Corporation.

190 B.L. Brooks et al. / Archives of Clinical Neuropsychology 25 (2010) 182 190 Korkman, M., Kirk, U., & Kemp, S. (2007c). NEPSY-II: Clinical and interpretive manual. San Antonio, TX: The Psychological Corporation. Palmer, B. W., Boone, K. B., Lesser, I. M., & Wohl, M. A. (1998). Base rates of impaired neuropsychological test performance among healthy older adults. Archives of Clinical Neuropsychology, 13 (6), 503 511. Schoenberg, M. R., Lange, R. T., & Saklofske, D. H. (2007). A proposed method to estimate premorbid full scale intelligence quotient (FSIQ) for the Canadian Wechsler Intelligence Scale for Children-Fourth Edition (WISC-IV) using demographic and combined estimation procedures. Journal of Clinical and Experimental Neuropsychology, 29 (8), 867 878. Schretlen, D. J., Testa, S. M., Winicki, J. M., Pearlson, G. D., & Gordon, B. (2008). Frequency and bases of abnormal performance by healthy adults on neuropsychological testing. Journal of the International Neuropsychological Society, 14 (3), 436 445. Thomas, S. V., Sukumaran, S., Lukose, N., George, A., & Sarma, P. S. (2007). Intellectual and language functions in children of mothers with epilepsy. Epilepsia, 48 (12), 2234 2240. van der Sluis, S., Willemsen, G., de Geus, E. J., Boomsma, D. I., & Posthuma, D. (2008). Gene-environment interaction in adults IQ scores: measures of past and present environment. Behavioral Genetics, 38 (4), 348 360.