Neuropsychological Assessment Battery (NAB): Introduction and Overview Travis White, PhD PAR, Inc. Acknowledgment Development of the NAB was made possible and funded in part by the following grants from the National Institute of Mental Health: • 1 R43 MH58501-01 • 2 R44 MH58501-02 Introduction • The NAB is a comprehensive, modular battery of 33 new neuropsychological tests, each with an equivalent form, developed to examine a wide array of cognitive skills and functions in adults, age 18 and older. • Decisions pertaining to content and format were guided by results of a national survey of neuropsychological assessment needs and practices (Stern & White, 2000), and by guidance from members of the NAB Advisory Council and other consultants. Rationale for the NAB • Arthur Benton (1992): The field of neuropsychology has lacked an integrated battery of instruments capable of providing highly sophisticated test data while requiring only a relatively brief administration time. • Oscar Parsons (1993): To meet current needs, such a battery should (a) have good psychometric characteristics, (b) include extensive normative and standardization data, (c) provide clinical information that satisfies a broad range of modern referral sources and questions, and (d) facilitate systematic research. Goal of Development • The goal underlying the development of the NAB was to address these needs by producing a new and innovative neuropsychological test battery that provides a comprehensive evaluation of neuropsychological functions in less than 4 hours. • The NAB incorporates the conceptual framework of Bauer (1994) and Tarter and Edwards (1986) by offering a separate Screening Module to indicate the need to administer additional domain-specific Modules. Screening Module Main Modules Screening Attention Domain Score Attention Module Screening Language Domain Score Language Module Screening Memory Domain Score Memory Module Screening Spatial Domain Score Screening Executive Functions Domain Score Spatial Module Executive Functions Module Flexibility • For those areas of functioning not included in the NAB (e.g., motor functioning, effort, mood/personality), the examiner can expand upon the NAB assessment with his or her favored instruments. • The individual examiner may choose to forego the Screening Module and administer any or all of the five domain-specific modules to a patient, based on specific clinical needs. • In addition, the flexibility inherent in the NAB also allows for selection of individual tests from each Module – rather than administering an entire domainspecific Module – when this type of non-battery focused assessment is clinically warranted. Survey • In order to ascertain the needs of the potential users of a new neuropsychological test battery, PAR conducted a comprehensive national Survey of Neuropsychological Assessment Needs (Stern & White, 2000). • The results served as a basis for the development of the NAB, vis-à-vis areas of functioning to include, length of battery, and other salient content and format characteristics of the battery. Survey • An important finding was the discrepancy between – The amount of time respondents thought was ideally needed for a comprehensive neuropsychological evaluation given current instrumentation (Mode = 5 to 6 hours; 25 % stated 4 hours or less) and – The amount of time they thought was required to conduct a realistic and reimbursable neuropsychological evaluation in today’s health care climate (Mode = 3 to 4 hours; 49% stated 4 hours or less). • 89% of respondents stated that there was no commercially available instrument that provided a comprehensive evaluation within the current time/funding constraints. Innovative Features of the NAB 1. Screening for both severely impaired and fully intact performance 2. Comprehensive coverage of functional domains 3. Combined strengths of flexible and fixed battery approaches to assessment 4. Avoidance of floor and ceiling effects 5. Reduced administration time 6. Coordinated norming (entire NAB normed on a single standardization group) 7. Demographically corrected norms based on age, education level, and sex 8. Provision of equivalent/alternate form 9. Increased user-friendliness for both examiner and examinee 10. Focus on ecological validity Dual-Screening Capability • Screening capability rated as moderate-to-very important by 74% of the survey respondents. • In practice, neuropsychological screening is typically geared toward identifying patients who show no signs of brain dysfunction and no need for extensive follow-up testing. This approach has been formally incorporated into two popular assessment instruments, the Dementia Rating Scale-2 and Cognistat (NCSE). Dual-Screening Capability • The NAB Screening Module provides screening recommendations at both ends of the ability spectrum. • For each NAB Screening Domain score, two recommendations are offered: (1) administer related module or (2) do not administer related module. Recommendations to forego the main module are made if: – the patient is fully intact (i.e., lacks impairment) and thus does not require administration of the analogous NAB Main Module because he/she would obtain similarly intact/above average scores – the patient is moderate-to-severely impaired and thus would not require administration of the analogous NAB Main Module because he/she would likely obtain similarly impaired scores • If the referral question requires greater quantification and description of the patient’s functioning, the user can always disregard the screening algorithm and administer the entire battery or the select functional Module(s). Comprehensive Coverage of Functional Domains • Reviews of the neuropsychological literature (e.g., Lezak, 1995; Mapou & Spector, 1995; Spreen & Strauss, 1998) have identified seven major functional domains: – Language and verbal communication functions – Spatial/perceptual skills – Sensorimotor functions – Attention and related information processing tasks (including working memory) – Learning and memory – Executive functions and problem-solving abilities – Personality, emotional, and adaptive functions. • This conceptual framework has been confirmed with factor analytic studies of various neuropsychological batteries (Larrabee & Curtiss, 1992; Leonberger et al., 1992). Comprehensive Coverage of Functional Domains • The NAB was developed with the overriding goal of providing a common set of core tests that serve as a reasonably comprehensive standard reference base suitable for most routine clinical applications. • Thus, the NAB is specifically not a “screening battery.” • The NAB is also not an exhaustive test battery that measures every conceivable neuropsychological skill and function. Comprehensive Coverage of Functional Domains • The survey of neuropsychologists directly guided the final content composition of the NAB into the following six modules: Screening, Attention, Language, Memory, Spatial, and Executive Functions. • Within each of the functional domains, results of the survey guided inclusion and exclusion of specific subdomains of assessment. Combined Strengths of Flexible and Fixed Battery Approaches to Assessment • The flexible and fixed battery approaches to neuropsychological assessment each have strengths and limitations. • In developing the NAB, we attempted to include as many strengths as possible, while avoiding as many weaknesses as possible. Combined Strengths of Flexible and Fixed Battery Approaches to Assessment Therefore, the NAB has the following characteristics: 1. Constant background of tests 2. Focused, patient-centered examination 3. Shorter administration times afforded by the efficient screening/test selection 4. Minimal reliance on clinical decision-making in test selection. 5. Standardized administration and scoring procedures across all tests 6. Quantitative summary indexes along with numerous measures of qualitative aspects of performance Avoidance of Floor and Ceiling Effects • Approximately 90% of survey respondents indicated that it would be moderately or very important for a new comprehensive test battery to be appropriate for high functioning examinees and should, therefore, avoid ceiling effects. • Approximately 73% of survey respondents indicated that a new battery should also be appropriate for severely impaired patients and should, therefore, avoid floor effects. Avoidance of Floor and Ceiling Effects • Thus, a guiding principle in the development of the NAB was the avoidance of both ceiling and floor effects, when appropriate. • For most tests in the NAB, a continuum of difficulty levels was included so as to provide a relatively normal distribution in test performance. • Difficulty ratings were provided by the Advisory Council members and used in the initial creation and selection of individual test items. • In addition, item difficulty statistics were calculated on field testing and standardization data to assure the adequacy of distributions. Reduced Administration Time • The NAB provides a reasonably comprehensive evaluation in a much briefer period than is currently available. • Approximately 71% of the survey respondents indicated that a realistic and reimbursable neuropsychological evaluation can be completed within 3-to-4 or 4-to-5 hours (excluding record review, interviewing, and report writing). • The entire NAB requires less than 4 hours to administer. In fact, in the Standardization sample, the majority of subjects completed the NAB in 2.5 to 3 hours. NAB Administration Time • • • • • • Screening Module = 45 min. Attention Module = 45 min. Language Module = 35 min. Memory Module = 45 min. Spatial Module = 25 min. Executive Functions Module = 30 min. • Full NAB (5 main modules) = 180 min. (3 hrs.) • Screening Module and Full NAB = 220 min. (3 hrs., 40 min.) Coordinated Norming • Whereas much is known about the psychometric properties of individual neuropsychological tests (Franzen, 1989; Lezak, 1995; Mitrushina, Boone, & D’Elia, 1998; Spreen & Strauss, 1998), very little effort has been devoted to the examination of how individual instruments function within a battery (Russell, 1994). • Given the fact that 85% of the survey respondents reported using a customized battery, the lack of psychometric data on customized batteries represents a very large gap in the neuropsychological knowledge base, and may lead to critical limitations in the overall validity of clinical decisions based on neuropsychological test data (Faust et al., 1991). Coordinated Norming • The NAB fills this critical gap by providing coordinated norms for all of the NAB tests and composite scores collected on the same standardization sample. • These coordinated norms allow for within- and between-patient score comparisons across the NAB. • Thus, the examiner can use a single set of normative tables (including the same age, education, and sex corrections) for the entire NAB, rather than dealing with the commonly used mixture of test-specific norms compiled in each examiner’s idiosyncratic “norms book.” Coordinated Norming • An important consideration in interpreting the performance of individual examinees is the magnitude of difference between planned comparisons of scores. • The coordinated norming of the NAB allows users to interpret score differences with two types of comparisons: – Statistical significance of score differences – Base rate of score differences Demographically Corrected Norms • The need to interpret neuropsychological tests within the context of an individual’s age, educational attainment, and sex has been well established in the field (c.f., Heaton, Grant, & Matthews, 1991). • Given that over 95% of the survey respondents viewed the availability of demographically corrected norms as moderately (18%) or very important (77%), the norms provided for the NAB represent a unique and critical feature. Demographically Corrected Norms • The NAB demographically corrected norm sample consists of 1,448 individuals. • Separate normative tables are provided for all combinations of the following demographic variables: – Age (18-29, 30-39, 40-49, 50-59, 60-64, 65-69, 70-74, 75-79, 80-97) – Education (<=11 years, 12 years, 13-15 years, >=16 years) – Sex Provision of Equivalent/Alternate Forms • An important aspect of neuropsychological assessment is the ability to monitor and document changes in functioning over time. • Survey results indicated that 96% of all respondents viewed the detection of change over time as a moderately (33%) or very important (63%) characteristic of a new comprehensive neuropsychological test battery. • Current neuropsychological instruments are poorly equipped to meet this goal because of a lack of equivalent, “repeatable” forms (Lezak, 1995) and a limited understanding of practice effects on neuropsychological testing (Sawrie et al., 1996). Provision of Equivalent/Alternate Forms These needs were addressed in two ways during development of the NAB: • Two parallel, equivalent forms were developed for each NAB module during the initial development phases. • Because many repeat testing sessions occur 6 months or more after the initial evaluation, a test-retest reliability study of the NAB was conducted using a 6month retest interval. Resulting SEMs and expected practice effects help differentiate meaningful score differences from artifactual practice effects. Increased User-Friendliness The NAB is more user-friendly than existing instruments with respect to: • Modularity • Portability • Face validity Modularity • Almost 90% of the survey respondents rated modularity as either moderately (29%) or very important (60%) for a new instrument. • Each of the six NAB modules are “selfcontained” and may be administered independently of the other modules. Portability • 76% of survey respondents rated portability as either moderately or very important. • NAB materials are highly portable because a minimal number of manipulatives are required and all necessary visual stimuli are integrated into a single stimulus booklet for each module. • All administration and scoring instructions are contained in the record forms, thus eliminating the need to juggle multiple forms and manuals during administration. • All materials necessary for an entire NAB administration fit into the provided attaché case. Portability • Approximately 73% of the survey respondents rated computerized administration as only slightly important or not at all important. • Although this finding is initially surprising, it is understandable because even laptop computers significantly reduce portability and raise design and psychometric problems. • Thus, the NAB is administered entirely by an examiner (i.e., not by computer). However, there is a computerized scoring software package (NAB-SP). Face Validity • Face validity is an important and often overlooked aspect of neuropsychological validation (Lezak, 1995; Nevo, 1985) • Face validity refers to whether a test appears to measure what it purports to measure, as perceived by: – Examinees who take it – Administrative personnel who decide upon its use – Other technically untrained observers, such as the examinee’s family (Anastasi & Urbina, 1997). • Tests that lack face validity are more prone to rejection by patients with brain dysfunction who are likely to be easily frustrated and fatigued. Face Validity • The face validity of the NAB was rated by the members of the Advisory Council, and items and tasks with poor face validity ratings were eliminated or modified. • Although the attractiveness of test materials is not often discussed in literature on face validity, the NAB includes modern, inviting, and colorful stimuli, materials, and artwork, including highquality digital photography. Focus on Ecological Validity • Ecological validity is the functional and predictive relationship between (a) performance on a set of neuropsychological tests during a highly structured, office-based test session and (b) behavior in a variety of real-world settings, such as home, work, or school (Long, 1996). • Over 79% of Survey respondents rated ecological validity as being either moderately or highly important attributes of a new comprehensive neuropsychological test battery. Focus on Ecological Validity • The development of the NAB specifically emphasized ecological validity. • For example, each NAB module (with the exception of Screening) includes one Daily Living test that is designed to be highly congruent with an analogous real-world behavior. • By definition, NAB Daily Living tests are multifactorial in nature. NAB Materials • Manuals – NAB Administration, Scoring, and Interpretation Manual – NAB Psychometric and Technical Manual – NAB Demographically Corrected Norms Manual – NAB U.S. Census-Matched Norms Manual NAB Materials • NAB Software Portfolio (NAB-SP) – Automates many steps involved in calculating raw scores and in obtaining normative scores and profiles. – Two score reports: Screening and Main Modules. – Choice of two normative samples. – Profile graphs, including overlays of multiple administrations. – Reports exportable to word processing programs. – Data exportable to spreadsheet and database programs. NAB Materials • Test Administration Materials – Record Forms: • One for each Module • One per each of two NAB equivalent forms • All necessary instructions for administration and scoring NAB Materials • Test Administration Materials – Response Booklets: • One each for Screening, Attention, Language, and Executive Functions Modules • One per each of two NAB equivalent forms • Used for tests that require the examinee to write, draw, or provide other similar responses NAB Materials • Test Administration Materials – Stimulus Books: • One for each Module • One per each of two NAB equivalent forms • Contain all visual stimuli presented to examinee for all tests other than Map Reading NAB Materials • Test Administration Materials – Manipulatives • Design Construction tests in both Screening and Spatial Modules use a set of five flat, blue plastic geometric shapes (Tans) based on the ancient Chinese puzzle game • Spatial Module Map NAB Materials • Test Administration Materials – Scoring Templates • Numbers & Letters tests in both Screening and Attention Modules require scoring templates to score omissions and commissions. General Principles Guiding the Development of the NAB • Tests must be easy to administer and score • Stimuli must be attractive and face valid • Total administration time for the five Main Modules must be 3 hours or less • Start with a large pool of items that represents a wide range of difficulty • Meaningful relationship between analogous Screening Module and Main Module tests • Theoretical foundation must combine empiricism (prediction) and cognitivism (constructs) • Test names should describe the content and/or procedures involved (“Dots” versus “Working Memory Test”) • Advisory council ratings must inform development activities Item Development/Reduction/Selection • For each test, at least two times the final number of items/stimuli were initially created (sometimes 10 times), using detailed development criteria, objective ratings (e.g., word frequency), and computerized manipulations. • Results of Advisory Council Ratings AND numerous field testing studies guided both item reduction/selection and equating of forms. Test/Item Characteristics Rated by Advisory Council • • • • • • Verbal encodability Clinical utility Difficulty Ecological validity Education bias Ethnic/racial/cultural bias • Sex bias • U.S. regional bias • Linguistic demands • Quality of stimuli, artwork • Stimulus satisfaction • Task appropriateness • Overall task satisfaction Standardization of the NAB Standardization Sites • Collection of the NAB standardization data started in September of 2001 and concluded in October of 2002. • NAB standardization data were collected at five sites that were selected to provide representation in each of the four geographic regions of the U.S. • Four of the sites were located at academic institutions with known expertise in neuropsychology; the publisher’s offices in Florida served as the fifth site. NAB Normative Samples The total NAB standardization sample consisted of 1,448 healthy, community dwelling participants, which formed the basis of the following normative samples: • Demographically corrected norms (N = 1,448) • Age-based, U.S. Census-matched norms (N = 950) Age-based, U.S. Census-matched Norms • The Age-based, U.S. Census-matched sample (N = 950) was abstracted from the total standardization sample. • Closely matches the characteristics of the current U.S. population with respect to education, sex, race/ethnicity, and geographic region. • Purpose: For making inferences regarding the adequacy of the tested ability in more absolute terms, i.e., compared to the population as a whole. Demographically Corrected Norms • Consists of 1,448 healthy, community-dwelling individuals ranging in age from 18 to 97 years. Of these 1,448 participants, 711 received Form 1 and 737 received Form 2 as part of the standardization study; no participant completed both NAB forms. • Purpose: For diagnostic inferences and for interpreting brain-behavior relationships, i.e., compared to age-, education-, and gender-matched peers. Demographically Corrected Norms • For inferring brain-behavior relationships, it has been well established in the neuropsychological literature that demographically corrected norms are the most appropriate normative standard (Heaton et al., 1993; Heaton et al., 1991; Lezak, 1995; Mitrushina et al., 1998; Spreen & Strauss, 1998). • The research literature has clearly established that performance on a neuropsychological test can be significantly impacted by an individual’s age, educational attainment, and sex, irrespective of potential brain dysfunction. • Thus, interpretation of brain-behavior relationships should be based on normative data either categorized according to different groupings of these demographic variables, or “corrected for” the effect of these variables. Demographically Corrected Norms • The demographically corrected norms are recommended for most situations encountered in clinical practice and, thus, they are the primary normative standard for the NAB. • All normed scores presented in the psychometric analyses and tables in the NAB manuals are based on the demographically corrected norms. Development of Norms • Selection of normative scores • Equating of Forms 1 and 2 using equipercentile equating methods • Verify the accuracy of the equating process • Conversion of raw scores to z scores • Examination of the effects of age, education, and sex • Continuous norming to create norm tables • Verify the accuracy of the demographic correction process • Calculate composite Index scores Selection of Normative Scores • The NAB consists of 33 individual tests, most of which provide at least several indicators of quantitative and qualitative performance. • Prior to beginning the norming process, all potential NAB scores were categorized into one of three types of scores: primary, secondary, or descriptive. • Several sources of information were used to categorize scores, including their (a) reliability, (b) presumed interpretive importance, and (c) content and construct validity. Types of Normative Scores Score Normative Metric Primary T scores (M = 50, SD = 10) Secondary Percentiles by age groups Descriptive Cumulative percentages for overall sample Equating of Forms 1 and 2 • Test equating refers to a family of statistical concepts and procedures that have been developed to adjust for differences in difficulty level on alternate test forms, thus allowing the forms to be used interchangeably. • Test equating adjusts for differences in difficulty between the two forms of a test, not for differences in content (Kolen & Brennen, 1995). Equating of Forms 1 and 2 • The equipercentile equating method was selected for use with the NAB because it is thought to have greater generalizability and applicability than mean and linear equating when test scores may deviate from a perfectly normal distribution (Kolen & Brennen, 1995), which is the case with many NAB scores. • Note that only NAB primary scores are equated. Secondary and descriptive scores are not equated by form; therefore, normative data for these scores are provided separately by form. Influence of Demographic Variables • Regression techniques were used to evaluate the potential effects of age, education, and sex on NAB raw scores. • Age, education, and sex were entered into separate regression equations as predictors, and the NAB primary z score was the dependent variable. The percentage variance in z scores (as reflected by the R2 value) accounted for by each demographic variable was recorded. • Next, the three demographic variables were entered into a stepwise regression equation to determine the impact on z scores of the demographic variables taken as a group. Continuous Norming • The method of continuous norming (Gorsuch, 1983b) was used to derive both the NAB demographically corrected norms and the age-based, U.S. Census-matched norms. • Continuous norming corrects for irregularities in (a) the distributions of scores within groupings of the norming variable and (b) trends in the means and standard deviations across groupings when group sample sizes are 200 or smaller (Angoff & Robertson, 1987). • Continuous norms provide a more accurate estimation of population parameters such as means and standard deviations because they are based on an equation that results from using all demographic groups, rather than only the one group for a particular table (Zachary & Gorsuch, 1985). Calculation of Index Scores • For each participant in the norm sample, the actual T scores on the tests that comprise the composite score were summed, and the cumulative frequency distribution of this new score was calculated. • The Module Indexes were scaled by converting the cumulative frequency distribution of the summed scores to a normalized standard score scale with a mean of 100 and a standard deviation of 15. • The Total NAB Index was calculated as the sum of the five Module Indexes, which results in each module contributing equally to the Total NAB Index, regardless of the number of tests that comprise individual Module Indexes. Composite Scores (M =100, SD = 15) Main Modules Screening Module • Attention Index • Screening Attention Domain • Language Index • Screening Language Domain • Memory Index • Screening Memory Domian • Spatial Index • Screening Spatial Domain • Executive Functions Index • Screening Executive Functions Domain • Total Screening Index • Total NAB Index Reliability and Score Differences Interrater Reliability • The consistency of agreement of test scores from rater to rater is also an important indication of a test’s reliability. This is especially the case for those subtests that require scorer judgment and decision-making. • The interrater reliability for the NAB was examined for the following subtests: Writing, Story Learning, Figure Drawing, Judgment, and Categories • Thirty Form 1 and 30 Form 2 standardization protocols were randomly selected and independently scored by experienced standardization examiners. • For the remaining tests, one-way single-measure intraclass correlation coefficients (ICC) were calculated Equivalent Forms Reliability • The equivalent forms reliability of the NAB was evaluated by applying generalizability theory (Brennan, 2001; Cronbach, Gleser, Nanda, & Rajaratnam, 1972; Shavelson & Webb, 1991). • In contrast to classical psychometric theory that posits true scores and a unitary or undifferentiated source of error, the application of generalizability theory allows the partitioning of various sources of variance using the familiar analysis of variance design. • Generalizability coefficients are considered analogues to traditional reliability estimates. • However, in contrast to the magnitude of traditional reliability estimates, generalizability coefficients of .60 or higher should be regarded as demonstrating very good reliability (Cicchetti & Sparrow, 1981; Mitchell, 1979). Equivalent Forms Reliability • Results showed good to excellent generalizability coefficients for most NAB primary scores. • The average percentage variance attributable to Form was 2.1% (Median = 0.4). • After the forms were equated, the average percentage residual variance attributable to form was 0.2% (Median = 0.0). Reliability of Composite Scores • Given that the alpha coefficient is an inappropriate estimate for many NAB scores, G coefficients were uniformly used as reliability estimates for the purpose of calculating the reliability estimates of Screening Domains, Total Screening Index, Module Index scores, and Total NAB Index scores. • The reliability coefficients for all composite scores were calculated with the formula recommended by Guilford (1954) and Nunnally (1978). Standard Errors of Measurement and Confidence Intervals • Standard errors of measurement are provided for all NAB primary test scores and composite Domain and Index scores using the following formula: SEM = SD * SQRT(1-rxx) • The SEM provides an estimate of the amount of error in an individual’s observed test score. • In this formula, SD is the standard deviation of the scores normative metric (i.e, 10 or 15) • 90% and 95% confidence intervals were developed for all NAB composite Domain and Index scores Score Differences • An important consideration in interpreting the performance of individual examinees is the magnitude of difference between planned comparisons of scores. • The coordinated norming of the NAB allows users to interpret score differences with two types of comparisons: – Statistical significance of score differences – Base rate of score differences Base Rate of Score Differences • The base rate of score differences addresses the actual occurrence rates, expressed as cumulative percentages, of score discrepancies that are present in the standardization sample. • It is quite possible to obtain difference scores that are statistically significant but that occur relatively frequently in the norm sample. • Base rates of score differences are provided for all Screening Module composite score pairs, and for all Main Module composite score pairs. Validity of the NAB Overview of Validity Studies • Evidence based on theory and test content (i.e., content validity) • Evidence based on internal structure – Intercorrelations of test and Index scores – Factor analyses (EFA and CFA) – Relationship between Screening Domain and Module Index scores • Evidence based on relationships to external variables • Evidence based on performance of clinical groups Content Validity • Reviews of the neuropsychological literature (e.g., Hebben & Milberg, 2002; Lezak, 1995; Mapou & Spector, 1995; Mitsrushina et al., 1998; Spreen & Strauss, 1998; Williamson et al., 1996) have identified seven major functional domains that are typically assessed: 1. 2. 3. 4. 5. 6. 7. Attention and information processing (including working memory) Language and verbal communication Spatial/perceptual skills Learning and memory Executive functions and problem-solving abilities Sensorimotor functions Personality, emotional, and adaptive functions. Content Validity • This conceptual framework has been confirmed with factor analytic studies of various neuropsychological batteries (Ardilla, Galeano, & Rosselli, 1998; Larrabee & Curtiss, 1992; Leonberger et al., 1992; Ponton, Gonzalez, Hernandez, Herrera, & Higareda, 2000) • This framework served as the underlying structure of the NAB. Content Validity • As described previously, results of the publisher’s survey of neuropsychological needs and practices guided the content composition of the NAB. • Those results provided strong support for organizing the NAB into a Screening Module and five main modules corresponding to functional domains: Attention Module, Language Module, Memory Module, Spatial Module, and Executive Functions Module. • Survey respondents reported a strong preference to continue using existing measures of sensorimotor functions and personality/emotional functions; that is, the preference was to not create new measures of these functions for a newly developed battery. Content Validity • Content validity deals with the issue of how well a group of items or tests is representative of the previously defined domain or domains of interest • Evidence of content validity is typically obtained by having knowledgeable experts examine the test material and make judgments about the appropriateness of each item and/or test and the overall content coverage of the domain. • In addition, content validity is often evaluated by examining the procedures and plans used in test construction. Content Validity Content validity of the NAB was established through a variety of methods, including: • Reviewing the neuropsychological assessment literature, including factor analytic work • Survey of neuropsychologists • Replicable development procedures for each NAB test • Extensive Advisory Council ratings and feedback Content Validity • The detailed procedures used to develop each NAB test are beyond the scope of this presentation. However, they are discussed at length in Chapter 2 of the NAB Psychometric and Technical Manual. • The methods used to create the NAB tests provide support for the content-related validity of each test and for the modular structure of the NAB. Internal Structure of the NAB • Intercorrelations among NAB scores (see Manual) • Exploratory factor analyses (EFA) • Confirmatory factor analyses (CFA) Exploratory Factor Analyses • Separate EFAs were performed for the primary scores of the Screening Module and the primary scores of the main modules. • The NAB standardization sample was used for the EFAs • Factors were extracted using principal axis factoring • Promax rotation of retained factors. • For both EFAs, three- to six-factor solutions were examined. • All factor solutions were interpreted using traditional methods (e.g., evaluation of scree plot and eigenvalues). • The theoretical underpinnings of the NAB and the meaningfulness of the constructs were interpreted according to the recommendations of Gorsuch (1983a, 1996) Summary of EFA Models • These EFAs were conducted as a means of forming additional hypotheses regarding the number and composition of the latent factors that underlie the observed NAB data. • The obtained EFA models show a fair degree of concordance with the conceptual model of neuropsychological constructs that underlies the NAB. • The hypotheses generated by the EFAs were evaluated in the construct-testing process of the subsequent confirmatory factor analyses (CFA). • CFA is intended as a theory or construct evaluation procedure (Stevens, 1996), and thus CFA results bear directly on establishing the validity of the NAB Domain and Index scores. Confirmatory Factor Analyses • CFAs were performed with the AMOS Version 4.0 structural equation modeling software program. • The NAB standardization sample was again used for the analyses. • A variety of fit statistics were evaluated. • The Comparative Fit Index (CFI) and the root mean square error of approximation (RMSEA) were given priority because they provide more stable and accurate estimates (Hu & Bentler, 1995). • CFI values at or above .95 suggest good fit. • RMSEA values at or below .06 suggest good fit. Summary of CFA Results • CFA of the Screening Module resulted in a five-factor model that mirrors the five functional neuropsychological domains purportedly measured by the Screening Domain scores. • CFA of the Main Modules resulted in a six-factor model that mirrors the five functional domains purportedly measured by the Module Index Scores, plus the presence of an additional psychomotor speed factor that underlies performance on several NAB tests. • CFA results support the construct validity of the Screening Domain and Module Index scores. Validity Evidence Based on Relationships to External Variables Healthy Validity Study • 50 Standardization subjects received a “gold standard” battery of neuropsychological tests within 10 days of their initial NAB examination. Clinical Validity Studies (~200 Patients) • • • • • • • Mild TBI (Full Clinical Battery): n = 31 Dementia (MMSE, DRS-2): n = 20 MS: n = 31 ADHD: n = 30 HIV/AIDS: n = 19 Aphasia (BNT, Token Test): n = 31 Inpatient Rehab (FIM): n = 39 Convergent Validity • Results of “Healthy Validity” study and Clinical Group studies were used to examine the convergent and divergent validity of individual NAB scores and indexes. • All validity coefficients are presented in the NAB Psychometric and Technical Manual. • The following selected data are illustrative of the overall convergent validity findings.