Repeatable Battery for the Assessment of Neuropsychological Status (RBANS): Preliminary Australian normative data

Abstract

The aim of the present study was to establish demographically adjusted normative data for the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) in an Australian context. RBANS data from 172 healthy Australian community dwelling adults enrolled in the “Using our Brains” (UoB) Brain Tissue Donor program between April 2002 and September 2005 were included in this study. The present study group differed from the original North American normative sample on all 12 subtests and five RBANS indices. Education and to a lesser degree age were related to RBANS performance. The present study provides complementary RBANS normative data for clinicians and researchers when interpreting RBANS results for Australian individuals closely matching those reported in this study.

Keywords:

• Australian normative data
• RBANS

Cognitive impairment is a hallmark feature of many neurological and psychiatric conditions (Dickerson et al., 2004; Wilk et al., 2002). Over the years various comprehensive test batteries have been devised in order to evaluate cognitive impairment across diagnostically diverse patient groups. Some of the better-known test batteries include the Halstead – Reitan Test Battery (Reitan & Wolfson, 1996), The Wechsler Memory Scale – revised edition (WMS-III) (Wechsler, 1997a) and Wechsler Adult Intelligence Scale – revised edition (WAIS-III) (Wechsler, 1997b).

Despite the popularity and proven clinical utility of these tests (Gontkovsky, Beatty, & Mold, 2004) they often create problems for clinicians and researchers assessing neuropsychological functioning in older adults, acutely ill inpatients and large research groups. Some of these problems relate to test difficulty and length, lack of portability, susceptibility to practice effects and the high level of expertise required to score, administer and interpret test findings (Ostrosky-Solis, Ardila & Rosselli, 1999; Randolph, Tierney, Mohr, & Chase, 1998; Wilk et al., 2004).

In order to minimise some of these problems a number of relatively brief cognitive screening tests have been developed, such as the Mini Mental Status Examination (MMSE) (Folstein, Folstein, & McHugh, 1975). But these measures are also not completely satisfactory.

Criticisms levelled at brief screening measures include a general insensitivity to milder forms of cognitive impairment (Peterson, Smith, Ivnik, Kokman, & Tangalos, 1994), a lack of appropriate normative data (Gontkovsky, Hillary, & Scott, 2002), susceptibility to ceiling effects in younger persons and a limited capacity to differentiate between cognitive domains to aid with differential diagnosis (Feher et al., 1992).

In 1998 Randolph devised a brief, highly portable test called Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) (Randolph, 1998) with extensive normative data (Aupperle, Beatty, Shelton, & Gontkovsky, 2002) capable of distinguishing between diverse cognitive domains in repeat format (Duff et al., 2005). While originally conceived of as a tool for the assessment of dementia in the elderly (Gontkovsky et al., 2004) its increasing popularity among clinicians has led to the examination of its clinical utility in diverse patient populations (Mooney et al., 2007). The RBANS has acceptable psychometric properties with respect to test – retest stability, internal consistency and concurrent validity (Gold, Iannone, & Buchanan, 1999; Hobart, Goldberg, Bartko, & Gold, 1999; Larson, Kirschner, Bode, Heinemann, & Goodman, 2005).

The effect of demographic variables such as age on test performance is well documented and for this reason it is commonplace for test manuals to provide additional age corrections to assist with interpretation of test findings (Carstairs, Myors, Shores & Fogarty, 2006; Collie, Shafiz-Antonacci, Maruff, Tyler, & Currie, 1999; Duff et al., 2003; Levine, Miller, Becker, Selness, & Cohen, 2004; Patton et al., 2005). In general, test performance deteriorates with increasing age. For example Geffen, Moar, O'Hanlon, Clarck, and Geffen (1990) found that performance of 153 healthy participants aged 16 – 86 years on the Rey Auditory Verbal Learning Test (RAVLT) decreased with increasing age. Similar results were obtained by Acevedo et al. (2000) following examination of category fluency task performance (animals & fruit) in 346 normal elderly subjects. Regardless of education, performance of participants in their 70s was 50% lower than that of individuals aged in their 50s.

The influence of education on neuropsychological performance has also been widely studied. Common to a number of these investigations is the finding that individuals with a higher educational level generally perform better than lower educated participants (Duff et al., 2003). For example in the Macquarie University Normative Neuropsychological Study (MUNNS) Carstairs and Shores (2000) reported significant educational effects in the aforementioned direction for Information, Vocabulary, Arithmetic, Comprehension, Similarities, Block Design, Object Assembly and Digit Symbol subtests of the WAIS-R for 399 community dwelling individuals aged 18 – 34 years. While the RBANS test manual provides corrections for age it does not make adjustments for education (Beatty, Mold, & Gontkovsky, 2003; Gontkovsky, Mold, & Beatty, 2002). However, three recent investigations of the RBANS in elderly participants have all found that education is a strong predictor of RBANS performance, accounting for up to 15% of the total variance in test scores (Gontkovsky, Mold et al., 2002; Lineweaver, Zone, Chelune, Hermann, & Dow, 2001). Consequently, the addition of adjustments for years of education has been recommended for more accurate interpretation of RBANS test results (Beatty et al., 2003).

The RBANS was developed and standardised on healthy North American volunteers. To the authors' knowledge there are no Australian RBANS normative data. However, previous research examining the performance of healthy Australian participants on tests devised and normed on North American samples indicates that Australian subjects, both children and adults, tend to perform differently on a wide range of neuropsychological measures (Casey & Heath, 1988; Shores & Carstairs, 2000), either similar or identical to tests that comprise the RBANS.

For instance, Hester, Kinsella, Ong, and Turner (2004) in a study of the Hopkins Verbal Learning Test in 203 community dwelling older Australians, found a generally lower level performance on the delayed memory indices relative to North American counterparts (10% less). Clark et al. (2004) also examined performance of Australian women, aged 56 – 67 years on executive function and working memory tasks. In general, Australian women performed slightly better than the age-matched US population for the Letter – Number Sequencing task and slightly worse for the Symbol Digit Modalities Test that resembles the coding test from the RBANS.

Holdnack, Lissner, Bowden, and McCarthy (2004) reported on normative data for the WAIS-III collected by Psychological Corporation of Australia and New Zealand (N = 297). Despite methodological limitations of the study precluding publication of research findings, higher performance was found among Australian participants relative to the US standardisation sample. Discrepancies are not limited to Australian-based individuals, in that other English-speaking countries also show differences relative to North American participants on measures contained within the RBANS. For example Crawford, Allan, and Gray (1995) found in favour of UK participants relative to the standardisation sample for the digit span subtest of the WAIS-III, despite being thought of widely as a culturally neutral task.

These cross-cultural differences have been thought to relate to a range of factors, which characterise a particular cultural group. These differences have been noted to remain even when demographic variables such as education and age are controlled for (Lucas et al., 2005). Such factors include differences in the demographic composition of the two normative populations being compared, the level and quality of educational systems, spoken language, socioeconomic status and home socialisation practices and values that combine to affect the relevance, familiarity and salience of test stimuli and test-taking attitudes (Shuttleworth-Edwards et al., 2004). As a result of cross-cultural differences, questions have arisen about the direct application of US normative data in Australian contexts. This has led Australian researchers and clinicians to call for Australian norms to help improve diagnostic accuracy (Clark et al., 2004; Hester et al., 2004; Holdnack et al., 2004; Shores & Carstairs, 2000).

Only one study to date has specifically examined cross-cultural influences on RBANS performance. Bluvshtein (2004) compared the results of 95 native Russian-speaking immigrants living in America on a Russian language adaptation of the RBANS. Results indicated no significant differences between the two groups on all RBANS indices with the exception of the Attention Index. However, the use of a convenience sample of individuals residing within the country in which the original RBANS population-based study was conducted limits the generalisability of findings to other cultural groups. Noting this, it is uncertain at this stage whether the RBANS is susceptible to cross-cultural differences common to many other neuropsychological tests.

Noting the above, the aims of the present study were threefold: (a) to evaluate the susceptibility of the RBANS to demographic influences, namely age and education; (b) to examine whether mean performances of Australian participants across the six RBANS summary scores are consistent with the hypothesised population mean of 100; and (c) to provide preliminary Australian norms for the RBANS, taking into consideration possible demographic influences on test performance.

Methods

Participants

Data for the present study were obtained through the “Using our Brains” (UoB) Brain Tissue Donor program between April 2002 and September 2005. This program is located in the Discipline of Pathology at the University of Sydney and invites members of the community to donate their brain tissue to medical research in life in much the same way as other organs are donated for transplant surgery. Participants were recruited through word of mouth, presentations at local community meetings and promotional activities. All were residing in the community at the time of enrolling in the program, predominantly in the Sydney metropolitan area.

Persons with a known or suspected history of neurological illness, psychiatric disturbance, substance abuse or any other condition that could affect cognitive performance were excluded from the analysis. The exclusion criteria adopted for this study were comparable to those used for the original US normative study. However, in contrast to the US sample, which underwent a nationwide population-based standardisation to provide a close correspondence between the normative sample and US Census proportions, the present group was not census matched.

Comparative differences between the present sample and Australian census matched data, where possible, are presented in Table I. In total, RBANS protocols from 172 participants were included in this study, 99 women and 73 men. Subjects were aged 20 – 89 years (M = 58, SD = 18), predominantly Caucasian (90%) with on average 14 years of formal education (SD = 4). Years of education were determined by the last full year of study completed. For example an individual who completed year 11 but went on to complete 2 years of full-time study at a tertiary institution received 13 years of education. Kindergarten was excluded from the calculation. Examination of the normative data presented in the RBANS manual would appear to suggest that the present sample was more highly educated and contained a higher percentage of women. Approval for the UoB program was obtained through the Human Research Ethics Committee at the University of Sydney and Sydney South West Area Health Service prior to the start of data collection.

Table I. Demographic characteristics compared to Australian 2001 census data

Age	n	% of Total	% of Total Australian Census Matched Statistics (2001)
20 – 39	37	22	40
40 – 49	13	8	20
50 – 59	33	19	17
60 – 69	31	18	11
70 – 79	43	25	8
80 – 89	15	8	4
Education (years)
8 – 9	14	8
10 – 11	26	15
12 – 13	28	16
14 – 15	26	15
16 – 17	34	20
18 – 19	16	9
20 – 21	12	7
21+	16	9
Gender (20 – 89)
Female	99	58	51
Male	73	42	49

Procedure

Participants completed the RBANS as part of the assessment process of the UoB program. The RBANS was administered along with a battery of other neuropsychological and clinical measures in the home of each donor. All tests were administered and scored according to the respective test manuals and entered into a secure database. The RBANS (Randolph, 1998) is an individually administered test that takes 20 – 30 min to complete and is designed to assess attention, language, visuospatial abilities and immediate and delayed memory.

It consists of 12 subtests, most of which are analogous to traditional neuropsychological measures that together yield five index scores and a total score. Normative information from the manual is based on 540 healthy individuals between the ages of 20 and 89 years.

Data analyses

All statistical analyses were carried out using SPSS 11.0 for Windows 2000 (SPSS, Chicago, IL, USA). Pearson's Product Moment Correlation Coefficients were calculated to determine the relationship between age, education and RBANS subtest and Index scores. A series of hierarchical regression analyses were conducted forcing the entry of age and education to determine the proportion of variance accounted for by these variables in each of the RBANS summary scores. The group means and standard deviations were calculated for all 12 RBANS subtests for each age and education subgroup. A series of one-sample t tests were used to determine whether mean performances of Australian participants across the six RBANS summary scores were consistent with the hypothesised population mean of 100.

Results

Correlation analyses yielded significant associations between education and five RBANS index scores, as well as between age and the Immediate Memory summary score (Table II). The relationship between subtest raw scores and age and education is also reported in Table II. Age was significantly correlated with 11 of the 12 subtests. Similarly, education was significantly correlated with nine of the 12 subtests.

Table II. Pearson's product moment correlations coefficients

	Age	Education
Indexes
Immediate Memory	−.16*	.27**
Visuoconstruction	−.04	.28**
Language	.06	.17*
Attention	−.10	.33**
Delayed Memory	−.07	.14
Total Scale Index	−.11	.33*
Subtests
List Learning	−.54**	.24**
Story Memory	−.39**	.26**
Figure Copy	−.29**	.17*
Line Orientation	−.18*	.23**
Picture Naming	−.19*	−.06
Semantic Fluency	−.19*	.19**
Digit Span	−.16	.29**
Coding	−.72**	.30**
List Recall	−.41**	.10
List Recognition	−.20**	.05
Story Recall	−.35**	.25**
Figure Recall	−.45**	.17*
Note: *Correlation significant at .05; **correlation significant at .01.

Results of the hierarchical regression analyses (Table III) indicate education to be the primary predictor of performance across all RBANS indices for the sample. The present sample performed significantly above the hypothesised normative group population mean on five of the six RBANS Index scores (Table IV). Table V presents the means and standard deviations for six age subgroups. These age bands were chosen to allow for a broad comparison with the original normative data. Table VI presents the means and standard deviations for the six age bands with <12 years or ≥12 years of education, given the influence of education in this study.

Table III. Hierarchical regression analysis using age and education to predict participant performance across RBANS index scores

RBANS Index	Primary predictor	B weights	Significance of B weight	Secondary predictor	B weights	Significance of B weight	Significance of r^2 for the equation
Immediate Memory	Education	.25	.001	Age	−.12	.10	.001
Visuoconstruction skills	Education	.28	.001	Age	−.001	.99	.001
Language	Education	.18	.02	Age	.08	.28	.05
Attention	Education	.32	.001	Age	−.06	.45	.001
Delayed Memory	Education	.13	.08	Age	−.06	.46	.15
Total Scale	Education	.33	.001	Age	−.07	.36	.001
Note: RBANS = Repeatable Battery for the Assessment of Neuropsychological Status.

Table IV. Comparison of mean performance of Australian participants across RBANS indices with the hypothesised normative population mean (i.e., 100)

Measure	Australian sample	US sample	Mean difference	t
Indexes
Immediate Memory	105	100	5.16	4.8**
Visuoconstruction	100.9	100	.92	.75
Language	114	100	14	15.9**
Attention	106	100	6	4.8**
Delayed Memory	102	100	2	2.1*
Total Scale	108	100	8	7.4**
Notes: RBANS = Repeatable Battery for the Assessment of Neuropsychological Status.
*t significant at .05; **t significant at .01.

Table V. RBANS subtests stratified by age

	Australian sample
	20 – 39 years	40 – 49 years	50 – 59 years	60 – 69 years	70 – 79 years	80 – 89 years
	(n = 37)	(n = 13)	(n = 33)	(n = 31)	(n = 43)	(n = 15)
Subtests
List Learning	32.9 (3.9)	32.6 (2.9)	30 (4.1)	29.2 (4.6)	26.7 (5.0)	24.4 (4.1)
Story Memory	20.1 (2.2)	20.8 (2.1)	18.8 (2.7)	18.9 (3.1)	17.6 (2.8)	15.7 (4.9)
Figure Copy	18.6 (1.6)	17.3 (2.6)	17.8 (2.7)	17.5 (3.0)	16.7 (2.8)	15.7 (3.4)
Line Orientation	18.4 (2.7)	18.2 (2.2)	17.4 (3.4)	18.4 (2.1)	17.6 (2.5)	15.9 (5.4)
Picture Naming	9.9 (0.2)	10 (0)	9.7 (.20)	9.9 (0.3)	9.9 (0.4)	9.8 (3.4)
Semantic Fluency	24.03 (5.8)	28.3 (5.0)	24 (5.8)	23.2 (6.0)	22.1 (5.3)	21.3 (4.5)
Digit Span	11.9 (2.7)	13.6 (2.2)	12.1 (2.3)	11.9 (2.5)	11.05 (3.0)	10.7 (2.7)
Coding	60 (7.6)	57.4 (7.2)	49.7 (6.8)	46.6 (7.1)	37.7 (12)	32.0 (10)
List Recall	8.3 (1.8)	8.07 (1.9)	6.3 (2.3)	6.8 (2.1)	5.8 (2.3)	5.4 (2.2)
List Recognition	19.8 (.4)	19.4 (1.0)	19 (2.8)	19.3 (1.2)	19.0 (1.4)	18.8 (1.4)
Story Recall	11.2 (0.8)	10.9 (1.1)	10.4 (2.3)	10.5 (1.6)	9.5 (1.9)	8.7 (2.3)
Figure Recall	16.1 (3.4)	14.2 (3.5)	13.7 (4.0)	13.2 (4.7)	12.0 (3.0)	9.2 (3.0)
	US sample
Measure	20 – 39 years	40 – 49 years	50 – 59 years	60 – 69 years	70 – 79 years	80 – 89 years
Subtests
List Learning	30.7 (4.3)	27.6 (4.4)	27.5 (4.7)	28 (4.5)	26.6 (5.0)	23.2 (4.5)
Story Memory	19.1 (3.3)	16.9 (3.2)	17.5 (3.7)	18.4 (3.5)	17.4 (3.6)	15.3 (3.9)
Figure Copy	19.1 (1.3)	18.3 (1.4)	18.2 (1.4)	18.1 (1.7)	17.8 (1.8)	17.3 (2.0)
Line Orientation	16.8 (3.0)	15.4 (3.0)	16.4 (2.9)	16.6 (2.9)	16.4 (2.8)	15.7 (2.6)
Picture Naming	9.6 (0.7)	9.4 (1.1)	9.4 (0.9)	9.7 (0.5)	9.6 (0.7)	9.1 (1.0)
Semantic Fluency	21.6 (3.7)	20.8 (5.0)	21 (5.0)	21 (4.6)	19.8 (5.2)	17.4 (3.7)
Digit Span	11.7 (2.5)	10.6 (2.2)	10.5 (2.4)	10.2 (2.1)	10.4 (2.5)	9.2 (2.2)
Coding	56.5 (8.8)	49.8 (8.1)	46.3 (8.9)	46.1 (7.9)	41.3 (9.0)	34 (6.8)
List Recall	7.5 (1.8)	6.3 (1.9)	6 (2.1)	6.0 (2.2)	4.9 (2.5)	3.9 (2.3)
List Recognition	19.8 (0.7)	19.7 (0.6)	19.5 (1.0)	19.4 (1.2)	19.2 (1.2)	18.8 (1.4)
Story Recall	10.1 (2.1)	8.9 (1.8)	9.1 (2.2)	9.3 (2.1)	9.0 (2.2)	7.4 (2.8)
Figure Recall	16.1 (2.9)	13.5 (3.3)	13.5 (3.3)	13.6 (4.0)	12.5 (4.2)	11.4 (4.1)
Notes: RBANS = Repeatable Battery for the Assessment of Neuropsychological Status.
US norms were reproduced for this paper with the permission of the author (Randolph, 1998).

Table VI. RBANS subtests for Australian participants stratified by age and education

Subtest	20 – 39		40 – 49		50 – 59		60 – 69		70 – 79		80 – 89
	<12 Edu	≥12 Edu	<12 Edu	≥12 Edu	<12 Edu	≥12 Edu	<12 Edu	≥12 Edu	<12 Edu	≥12 Edu	<12 Edu	≥12 Edu
List Learning	29.6 (4.0)	33 (4.0)	33 (3.7)	32.5 (3.2)	29.4 (5.0)	30.8 (3.2)	28.3 (5.2)	30.1 (3.3)	26.5 (6.4)	25.4 (4.6)	23.7 (5.5)	25.6 (3.2)
Story Memory	19.8 (1.3)	20 (2.3)	21.5 (1.7)	20.7 (2.3)	17.8 (3.0)	18.8 (2.6)	19.2 (3.8)	19 (2.3)	16.9 (3.5)	17.4 (2.6)	14.7 (5.7)	16.9 (4.3)
Figure Copy	17.8 (1.4)	18.6 (1.7)	15.7 (2.9)	17.5 (12)	18.6 (1.9)	17.8 (2.9)	17 (3.3)	17.9 (2.3)	16 (2.8)	16.9 (2.7)	16.7 (2.2)	15.4 (3.8)
Line Orientation	19.1 (.93)	18.4 (2.8)	16.8 (3.3)	18.0 (2.4)	18 (1.7)	17.3 (4.0)	18 (2.7)	19.2 (1.1)	17.3 (2.5)	17.7 (2.5)	14 (6.5)	18.8 (1.0)
Picture Naming	10 (0)	9.9 (.2)	10 (0)	10 (0)	10 (0)	9.9 (.21)	10 (0)	9.9 (.32)	9.9 (.30)	9.8 (.41)	9.9 (.11)	9.9 (.35)
Semantic Fluency	23.7 (4.9)	24 (6.0)	25 (5.6)	28.5 (4.9)	23 (6.0)	24 (5.9)	23 (6.6)	23.3 (4.8)	22.4 (6.2)	22 (5.1)	19 (2.3)	22.8 (5.5)
Digit Span	12 (2.4)	11.8 (2.7)	13.3 (3.1)	13.8 (1.7)	12.2 (1.9)	12.1 (2.4)	11.6 (2.4)	12.1 (2.3)	9.7 (1.8)	11.5 (3.2)	10.4 (2.5)	11.5 (3.0)
Coding	58.6 (7.7)	59.6 (7.6)	53 (5.4)	57 (7.0)	45 (4.0)	51 (6.2)	46.1 (8.9)	49 (5.4)	35 (11.6)	38.7 (12)	24.9 (8.4)	38 (7.9)
List Recall	7.3 (1.4)	8.3 (1.8)	7.8 (2.7)	8.2 (1.8)	5.4 (1.9)	6.8 (2.2)	6.9 (2.3)	6.8 (2.1)	6.8 (2.1)	5.5 (2.3)	5.1 (2.8)	6 (1.9)
List Recognition	19.8 (.67)	19.8 (.53)	19.3 (.96)	19.4 (1.0)	19.5 (.52)	18.8 (3.3)	19.4 (1.3)	19.3 (1.2)	19.4 (.81)	18.8 (1.6)	18.7 (.95)	18.8 (1.8)
Story Recall	11.2 (.44)	11.2 (.86)	10.8 (.96)	10.8 (1.2)	9.6 (2.0)	11 (2.4)	10.6 (1.5)	10.4 (1.7)	9 (2.1)	9.7 (1.8)	7.6 (2.9)	9.3 (1.6)
Figure Recall	13.6 (4.4)	16.3 (3.4)	12 (3.9)	14 (3.7)	14 (4.1)	13.6 (.82)	12.2 (4.9)	14.3 (4.3)	12.1 (3.3)	11.9 (3.5)	10.3 (2.8)	9.4 (.11)
	(n = 3)	(n = 34)	(n = 2)	(n = 12)	(n = 9)	(n = 24)	(n = 9)	(n = 22)	(n = 11)	(n = 30)	(n = 6)	(n = 10)
Notes: RBANS = Repeatable Battery for the Assessment of Neuropsychological Status.

Discussion

The ability of the RBANS to evaluate discrete cognitive abilities in brief format makes it a valuable tool for researchers and clinicians. To our knowledge there have been no published Australian RBANS norms despite evidence of cross-cultural differences on similar and identical tasks that comprise the RBANS. The aim of the present study was to evaluate the effect of demographic variables on RBANS performance in an Australian cohort. An additional aim was to examine possible differences between the present findings and the original US standardisation sample and establish Australian norms for the RBANS.

As detailed in the introduction, performance across a number of cognitive tasks has been shown to decrease, with age. Consistent with this, evidence from a variety of sources has shown that the various RBANS measures are also sensitive to the effects of age (Beatty et al., 2003). In the present study 11 out of 12 RBANS subtests and one of the six index scores correlated significantly with age. There was considerable difference between the different age groups in the present study particularly at the age extremes.

Education is known to affect neuropsychological performance in healthy adults (Patton et al., 2003). That is, individuals with fewer years of education tend to perform less well on a range of cognitive tasks relative to age-matched peers who have attained more years of formal education.

Recently, a number of articles have criticised the RBANS for its lack of education-based norms given its demonstrated sensitivity to education (Larson et al., 2005). Consistent with previous research, education was a primary predictor of performance on each RBANS index score for this sample (Lineweaver et al., 2001). The number of years of education was the greatest predictor of performance scores for the RBANS Immediate Memory, Visuoconstruction and Total score indexes, accounting for 7%, 8% and 11% of the total variance respectively. Education correlated significantly with five of the six RBANS index scores and nine of the 12 RBANS subtests. Consequently, the current dataset was further stratified by years of education to help reduce possible problems with interpretation using only age corrections. For example, using age-stratified norms, the mean performance of 80 – 89-year-old individuals on the coding task was 32. However, when results were further stratified by <12 years of education, the mean performance fell to 24.9 compared with a rise to 38 for individuals with ≥12 years of formal education.

Much of the normative data accompanying neuropsychological tests is based on North American samples. However, for a number of these measures performance of healthy North American participants has been shown to differ from that of other English-speaking populations including Australian samples (Collie et al., 1999).

These differences raise serious questions about the validity of directly applying such normative data in Australian contexts. Consistent with previous research, examination of the original North American normative data and that of the Australian participants in the present study showed that the two groups performed similarly but not identically. Australian subjects performed above the hypothesised US normative group mean on all RBANS index scores with the largest difference found on language-based tasks.

Results were more variable for individual RBANS subtests. While at this stage the small sample size prevented statistical testing of individual subtest differences, using the age bands provided by the authors of the RBANS, Australian participants performed slightly better on all verbal memory tasks (the exception being the list recognition subtest), the line orientation, picture naming, semantic fluency and digit span subtests. As to be expected given the findings above, the largest difference between the two groups occurred on the semantic fluency subtest, where Australian participants scored up to 8 points higher.

While it is possible that the discrepancy in performance between the groups, particularly on language-based tasks, relate to real cross-cultural differences, it is also possible that the differences observed represent differences between the groups in terms of demographic characteristics.

Particularly important in the present context is previous research that has consistently demonstrated gender differences by healthy participants (Beatty et al., 2003) across a number of neuropsychological measures. For example, female subjects often perform better than male subjects on verbal fluency tasks (Benton & Hamsher, 1989), whereas male subjects by contrast tend to perform better on measures that are heavily dependent on visual processing (Kimura, 1999). It is possible that differences observed between the two groups reflect the higher level of education and a greater proportion of women in the present study. This is consistent with the only investigation of gender differences in RBANS performance that found that female participants obtained significantly higher results on the language index (Beatty et al., 2003).

The present results should be interpreted in light of a number of methodological constraints. It is important to emphasise that these data were collected predominantly from highly educated middle-aged Caucasians living in the Sydney metropolitan area. The ability to generalise these results to more diverse cultural groups and educational backgrounds will need to be assessed in future studies. Due to a restricted sample size, the present study was unable at this stage to investigate gender differences. It is notable that Beatty et al. (2003) reported gender effects with the RBANS for 278 men and 353 elderly North American women residing in the community.

A larger sample size would also help to strengthen the present results because small cell frequencies were obtained for a number of the age groups particularly when stratified further by years of education, preventing statistical analysis of differences between the two groups at a subtest level.

The present study extends the normative data available for the RBANS to include education adjustments and Australian participants. It is not our intention to replace the normative data contained within the test manual. Rather, we offer an additional resource for clinicians and researchers, to use with Australian individuals closely resembling those represented in this sample.