See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/247492549

Development and psychometric properties of the Brief Test of Attention

Article in The Clinical Neuropsychologist · January 1996

DOI: 10.1080/13854049608406666

CITATIONS READS

124 5,471

3 authors, including:

David Schretlen Jason Brandt

Johns Hopkins Medicine Johns Hopkins University School of Medicine
233 PUBLICATIONS 10,421 CITATIONS 407 PUBLICATIONS 26,183 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Dementia Risk Assessment View project

Neural bases and psychometric approaches to decision-making. View project

All content following this page was uploaded by David Schretlen on 21 May 2014.

The user has requested enhancement of the downloaded file.

The Clinical Neuropsychologist 0920-1637/96/1 00 1-80$12.00
1996, Vol. 10, No.1, pp. 80-89 @ Swets & Zeitlinger

Development and Psychometric Properties of the Brief

Test of Attention *
David Schretlenl, Julie Hoida Bobholz2, and Jason Brandt3
IJohns Hopkins University School of Medicine, 2Chicago Medical School, and 3Johns Hopkins University
School of Medicine

ABSTRACT

The development and preliminary psychometric properties of a new instrument called the Brief Test of
Attention (BTA) are described. In terms of the conceptual model proposed by Cooley and Morris (1990),
the BTA is best described as a measure of auditory divided attention. The test consists of two parallel forms
that require less than 5 minutes each to administer and score. The results of testing 926 patients and normal
control subjects support the BTA's reliability, equivalence of forms, and construct validity. Coefficients
alpha for the entire test range from .82 to .91, while between-form correlations range from .69 to .81.
Neither practice nor interference effects were found to influence performance from the first to second form
administered. Correlation and principal components analyses showed that the BTA correlates more
strongly with widely accepted measures of attention than with other cognitive tasks, and more strongly
with complex than simple attention tasks.

Impairments of attention characterize most neu- (1985), captured scientific attention long after
ropsychiatric disorders. Attentional deficits are most attention tests enjoyed a secure place in
central to attention-deficit hyperactivity disorder clinical neuropsychology. Few experts would
(Barkley, Grodzinsky, & DuPaul, 1992) and de- question the clinical utility of the Digit Span and
lirium (Mesulam, 1985), but also have been as- Digit Symbol (Wechsler, 1939), Trail Making
sociated with schizophrenia (Braff, 1993), de- (Reitan, 1958), Stroop Color-Word (Stroop,
pression (Hartlage, Alloy, Vasquez, & Dykman, 1935), or continuous performance tests. How-
j 1993), traumatic brain injury (Stuss & Gow, ever, aside from Mirsky's (1989) factor analytic
1992), learning disabilities (Fleisher, Soodak, & approach to defining components of attention,
Jelin, 1984), cortical and subcortical dementia clinical tests rarely are described in terms of
syndromes (Cummings & Benson, 1992), epi- these conceptual models.
lepsy (Mirsky, 1989), and alcohol abuse (Oscar- One possible explanation for this dissociation
Berman & Bonner, 1985), to name just a few. between clinical practice and experimental neu-
I The majority of widely used clinical tests of ropsychology is that accepted clinical measures
attention were developed without respect to the- mix different components of attention, and con-
ories of attention. Elegant cognitive and neuro- found attention with other, often complex cogni-
anatomical models of attention, such as those tive processes. Cooley and Morris (1990) de-
scribed a framework for conceptualizing the ta~k
. proposed by Triesman (1961), Atkinson and
Shiffrin (1968), Gibson (1969), Posner and
ISnyder (1975), Wickens (1984), and Mesulam
demands of various clinically accepted measures
of attention. They argue that both sustained and

* The Brief Test of Attention@ may be obtained by contacting the first author. The authors wish to thank two
anonymous reviewers for their helpful comments on an earlier draft of this manuscript. Address correspondence
to: David Schretlen, Ph.D., Johns Hopkins Hospital, 600 N. Wolfe St., Meyer 218, Baltimore, MD 21287-7218,
tel: (410) 955-3268, USA.
Accepted for publication: April 26, 1995.
 THE BRIEF TEST OF ATTENTION 81

divided attention tasks can be conceptualized as The Brief Test of Attention consists of two
special cases of a basic selective attention pro- parallel forms that are presented via audio cas-
cess. Every selective attention task involves two sette. Subjects are administered both forms;
components: target identification (attend) and each requires 4 minutes to administer and score.
distractor inhibition (inhibit). According to this On Form N, a voice reads 10 lists ofletters and
model, sustained attention tasks are conceptual- numbers (e.g., "M-6-3-R-2") that increase in
ized as the extension of these two component length from 4 to 18 items. The subject's task is
processes over time. Divided attention is under- to disregard the letters and count how many
stood as requiring the performance of two simul- numbers are read aloud. Each list is followed by
taneous selective attention tasks. Cooley and 5s of silence, during which the subject reports
Morris conceptualize factors that influence the how many numbers were recited. The same 10
"inhibit" and "attend" components in terms of lists are presented as Form L, for which the sub-
four levels of processing. At the most basic ject's task is to disregard the numbers and count
level, tonic arousal regulates general informa- how many letters are read aloud. Unlike digit
tion processing (Mesulam, 1985). Modality- span tests, the subject is not asked to recall
linked sensory registration processes, as re- which numbers (or letters) are presented. The
quired by the detection of light flashes or number of correctly monitored lists is summed
buzzes, define the second level of processing. across forms; thus, total BTA scores can range
Modality-specific perceptual processes, such as from 0 to 20.
matching designs or letters, define the third The BTA was tape recorded in a sound-atten-
level. The fourth level involves conceptual pro- uated studio. The voice is that of a female radio
cessing. Relevant tasks typically involve the use broadcast professional who read test stimuli at
of multiple cognitive processing systems, cross- the rate of one letter or number per second. Each
modal comparisons, and/or the allocation of pro- trial is followed by 5s of silence. After all 10
cesses between systems. Finally, Cooley and lists were recorded, the resulting production was
Morris argue that each of the above levels of re-mastered using digital audio recording to fil-
processing may be represented within five (ver- ter out noise between trials. The first 10 trials
bal, spatial, memory, motor, and executive) are introduced on tape as, "Brief Test of Atten-
functional neuropsychological systems. tion, Part A." However, the same 10 trials are
In this article, we use the model proposed by duplicated on the tape and subsequently intro-
Cooley and Morris (1990) to describe a new duced as, "Brief Test of Attention, Part B." In
measure of auditory divided attention called the this way, either form N or L can be presented
Brief Test of Attention (Schretlen, 1989). We first. The effect of order of presentation is dis-
examine its psychometric properties, and com- cussed below. Prior to beginning the tape, direc-
pare its processing requirements to those of tions are read to the subject by the examiner.
other tests of attention. The examiner also reads two short sample lists.
If the respondent fails all three trials of both
sample lists, the test is discontinued; otherwise,
BRIEF TEST OF ATTENTION the entire form is administered. The tape is
never stopped during the administration of a
Development of the Brief Test of Attention given form. However, in some circumstances,
(BTA) began in 1989. The primary aim was to the use of only one form will yield reliably in-
devise a brief, relatively simple and easily ad- terpretable results.
ministered test of auditory divided attention that
would be sensitive to subtle attentional impair-
ments. A secondary goal was to reduce con-
founding task demands such as psychomotor
speed or conceptual reasoning.
82 DAVID SCHRETLEN ET AL.

STUDY 1: RELIABILITY AND were administered the BTA as part of their examina-
DEMOGRAPHIC CORRELATIONS tions. The BTA was not administered in a systematic
order relative to other tests. Although demographic
information was obtained for most subjects, the avail-
The purpose of this study was to document basic able psychometric data varied greatly, depending on
psychometric properties of the Brief Test of At- research protocols and patient referral questions.
tention in normal and clinical samples.

RESULTS
METHOD
Internal Consistency
Subjects Based on the 349 normal adults and children,
internal consistency analyses yielded a coeffi-
Normal sample
cient alpha of .82 for the BTA (.71 for Form L
This sample consisted of 275 adults and 74 children
who served as normal control subjects in one of five and .66 for Form N). When these data, were
separate studies. The children included all second (n = pooled with those of 480 patients for whom item
24), fifth (n = 25), and eighth (n = 25) grade students scores were recorded, and internal consistency
of a local elementary school. Although the elementary analyses were repeated using this combined
(n = 74) and college (n = 62) students, as well as 45
sample (n = 829), the coefficient alpha increased
adult normal control subjects did not undergo any
particular screening procedures, the remaining (n = to .91 for the BTA (.83 for Form Land .84 for
213) adult subjects were screened for dementia, severe Form N).
psychiatric disorders, and current substance abuse.
Form Equivalence
Clinical sample
Based on the 349 normal adults and children, the
This sample consisted of 577 patients drawn from
studies and clinics conducted at the Johns Hopkins Pearson correlation (r) between forms Nand L
University and Hospital. Included were patients with was .69. However, the between-forms correla-
schizophrenic (n = 55), affective (n = 105), sexual (n tion increased to .81 for the combined (n = 926)
= 24), substance abuse (n = 16), eating (n = 14), men- normal and clinical samples. Despite the modest
tal retardation (n = 16), and other (n = 40) psychiatric
disorders, as well as Huntington's disease (n = 56), correlation between forms, paired-sample t-tests
dementia (n = 32), traumatic brain injury (n = 44) and revealed that normal adults produced virtually
adrenoleukodystrophy(n = 51). The primary diagno- identical scores on Forms Land N (8.5, SD =
sis was either unavailable or not recorded for 124 pa- 1.6 vs. 8.5, SD = 1.5; t(274)= 0.67; p = .50). The
tients. Demographic characteristics of both normal patients produced lower and more variable
and patient samples are shown in Table 1.
scores overall, but they also showed virtually
Procedure identical performance on Forms Land N (5.6,
After giving voluntary informed consent, control sub- SD = 2.9 vs. 5.5, SD = 2.9; t(596)= 1.02; p = .31),
jects were administered the BTA along with whatever
other measures were included in the protocol in which
as did the normal children (t(73) = -0.62; P =
.54).
they served. Patients referred for clinical evaluations

Table 1. Demographic Characteristics of Normal and Clinical Samples.

Demographic Age Sex Race Education

Variable n M (SD) M/F W/B/O' M (SDi

Normal Samples
Adults 275 47.1 (19.3) 112/163 205/30103 14.4 (2.7)
Children 74 10.1 (2.5) 38/36 70/410 4.0 (2.5)
Clinical Sample 577 40.4 (15.7) 327/250 336/201/73 12.2 (3.2)4

'Race: W = White; B = Black; a = Other. 2Education expressed in years completed. 3Race was not recorded for
40 normalcontrolsubjectsand33 patients.4Educationnotrecordedfor 37 patients.
 THE BRIEF TEST OF ATTENTION 83

The effect of order of administration was ex- covariance, the main effect of sex disappeared
amined by subtracting each subject's score on (F(l.272) = 1.12; p = .29). Among the normal
whichever form was administered first from that adult sample were 30 African-Americans who
obtained on the form administered second. A were significantly older (t(194) = -4.36; p < .001)
practice effect should result in positive differ- and less educated (t(1SS)= 2.72; p < .01) than
ence scores; an interference effect should lead to White subjects in the same age range. Neverthe-
negative difference scores. However, if intra- less, White and Black subjects did not produce
individual variability across forms is random, significantly different BTA scores (17.2, SD =
then the difference scores should be normally 2.5 vs. 16.3, SD = 3.2: t(194)= 1.68; p = .094).
distributed with a central tendency that ap- When BTA scores were regressed on age, sex,
proaches zero. In fact, the mean difference score race, and education using stepwise multiple re-
of the 765 patients and normal subjects for gression, only age accounted for significant
whom order was recorded did approach"zero (M variance in the BTA performance of normal
= .15, SD = 1.8 points). The distribution of dif- adults (Adjusted R2 = .085; p < .0001).
ference scores was mildly leptokurtic (.30), indi- Cumulative frequency data for the BTA are
cating that more subjects than expected pro- presented in Table 2.
duced difference scores of zero. The distribution Within the sample of 577 psychiatric patients,
showed negligible skewness (-.07), suggesting BTAscorescorrelatedwithboth age(r =-.36, p
that exposure to the first form given neither fa- < .001) and education (r = .32, p < .001). Male
cilitated nor impeded performance on the second and female patients did not differ in their total
form. In fact, over 97% of normal subjects and BTA scores (11.4, SD = 5.4 vs. 10.6, SD = 5.7,
93% of patients produced scores that differed by respectively; t(595)= 1.81; p = .07). Total BTA
<;3 points between Forms Land N. scores were regressed on age, sex, education,
and race using stepwise multiple regression with
Impact of Demographic Characteristics forward variable selection. Three predictors
Among the 74 second, fifth, and eighth grade each accounted for significant unique variance
children, BTA scores were highly correlated in BTA scores, as demonstrated by the incre-
with both age (r = .65) and education (r = .62), mental adjusted R2S associated with each: age
which were very highly correlated with one an- (.141), education (.114), and race (.037). The
other (r = .98). The 38 boys produced signifi- final model was associated with an overall ad-
cantly lower BTA scores than the 36 girls (11.2, justed R2 of .292, which was highly significant
SD = 3.9 vs. 13.3, SD = 3.8: ten) = -2.37; p = (F(3.526)= 73.6; p < .0001).Educationand race
.021). In order to examine further the impact of were confounded in the clinical sample, but AN-
demographic characteristics on test perfor- CaVA revealed that the main effect of race
mance, BTA scores were regressed on age, sex, (F(l,549) = 23.0; p < .001) remained significant
and education (race was ignored because only 4 after age, sex, and education were entered as
subjects were not White). Using stepwise multi- covariates. When patients were grouped accord-
ple regression with forward variable selection, ing to years of education ( < 9, 9-11, 12, 13-16),
age and sex each met entry criteria for inclusion examination of BTA scores revealed that Black
in a highly significant model (Adiusted R2 = .45; patients scored 1.5 to 2.0 points lower than
p < .0001). White patients in each of the three upper educa-
In the sample of normal adults, both age (r = tion subgroups, although race-related diffe-
-.32, p < .001) and education (r = .13, P < .05) rences were not found among patients with less
were correlated with BTA scores. Women sco- than 9 years of education (Blacks, 7.6, SD = 4.8;
red marginally higher than men (t(273) = -2.18; p Whites, 7.2, SD = 5.8; p = .79).
= .03), but the men were nearly 10 years older Only 3 (1%) of the normal adults scored less
on average, a difference that was highly signifi- than 4 on either Form L or Form N. Based on
cant (t(273)= 3.86; p < .001). When age was en- this finding, a series of frequency analyses was
tered as the covariate in an analysis of conducted using all clinical cases for whom the
84 DAVID SCHRETLEN ET AL.

Table 2. Percent Cumulative Frequency of BTA Scores for Normal Control Subjects by Age Group.

Age range 6-8 9-11 12-14 17-19 20-39 40-59 60-69 70-81
n =24 n = 25 n = 25 n = 24 n = 89 n = 54 n = 68 n=40
BTA
Score

20 100 100 100 100 100 100

19 96 67 75 82 82 93
18 100 84 50 55 67 68 78
17 96 84 46 40 44 60 65
16 88 72 29 23 35 47 55
15 76 68 17 15 26 38 40
14 100 72 36 4 9 17 24 35
13 96 52 24 3 13 16 30
12 88 44 20 9 9 23
11 75 36 12 4 4 15
10 63 28 8 2 3 15
9 58 12 8 3 10
8 50 8 4 2 5
7 33 8 3
6 21
5 13
4 8
3 4
2 4
1 4
0 4

order of test administration was recorded (n = form given produced an abnormal total BTA
548). First, patients who scored < 4/10 on the score.
first form given were identified. Their protocols
then were examined to determine how many of
these patients produced abnormal scores (i.e., DISCUSSION
more than two SDs below the normal age group
mean) on the total test. Altogether, 142 (98 %) of The results of this study indicate that the Brief
the 145 patients who scored < 4 on the first form Test of Attention has adequate internal consis-
given also produced abnormal total BTA scores. tency. Correlations between Forms Nand L ran-
The remaining 3 patients all were '275 years old. ged from .69 to .81. Between-form difference
Conversely, when those patients who scored scores were normally distributed with a near-
> 7/1 0 on the first form given were identified (n zero mean, and were <;3 points for 93-97% of
= 155), and their total BTA scores were exam- subjects. Among normal adults, age alone ac-
ined, not one produced an abnormal total BTA counted for significant unique variance in BTA
score. As shown in Table 1, performance on the performance, although age and sex both contrib-
BTA is highly age-dependenCFor those who are uted significantly to the BTA performance of
70 or older, total scores of '28120 are within nor- children. Among the more heterogeneous pa-
mal limits. Thus, any elderly subject who scores tients, age, education, and race all made signifi-
>7/l 0 on the first form given thereby demon- cant contributions to the final regression model.
strates normal performance. However, the pres- Sex did not account for significant variance in
ent analyses indicate that, regardless of age, no BTA performance in either adult sample. Over-
patient who earned> 7/1 0 points on the first all, the BTA score frequency distributions ob-
 THE BRIEF TEST OF ATTENTION 85

served in these normal subjects must be viewed more highly with backward than forward digit
as preliminary, as they are based on samples of span.
convenience. Cutoff scores used to identify Part A of the Trail Making Test (Reitan,
attentional impairment might need to be lowered 1958) can be viewed as a selective attention task
by 1-2 points for persons with less than 12 years in which the targets are presented visually and
of schooling, but verification of this will require output is motoric. Part B of this test is better
further study. Further study also is needed to described as requiring divided attention because
establish test-retest reliability of the BTA; such it involves two simultaneous selective attention
research is in progress. tasks (i.e., monitoring two series to identify tar-
gets). Consequently, we predicted that the BTA
would correlate more highly with Part B than
STUDY 2: CONSTRUCT VALIDITY Part A of this test.
Word reading and color naming portions of
An analysis of the BTA based on the model of the Stroop paradigm (Stroop, 1935) both require
Cooley and Morris (1990) clearly places it simple selective attention at the perceptual level
within the verbal/linguistic functional system. of processing, but also require involvement of
On Form L, the subject listens to alpha-numeric both visual and verbal functional systems. The
lists and must detect letters (attend) while ignor- interference portion of the Stroop involves nam-
ing numbers (inhibit). This selective attention ing the color ink in which (different) color
process is extended to a lesser degree (the lon- words are printed. This can be conceptualized as
gest list is 18 s) than is the complexity of simul- requiring either a form of selectivity (i.e., inhib-
taneous processing. That is, after each target is iting the bias to read color words), or as requir-
detected, the subject must reallocate attention to ing divided attention (i.e., perceiving both the
the cumulative sum of previously reported let- color ink and color word, then selecting the for-
ters and increase it by one. Therefore, the BTA mer). In either case, we predicted that BTA per-
is conceptualized as an auditory perception task formance would correlate more highly with the
that requires divided attention to a much greater interference trial than with color naming or
extent than sustained attention. Both the percep- word reading.
tual (i.e., distinguishing letters from numbers)
and conceptual (counting from 1 to 12) require-
ments of the BTA are simple. The BTA is de- METHOD
signed to reveal impairments of the ability to
divide attentional resources between these si- Subjects and Procedure
multaneous tasks. Implicit in the model is the Subsamples for this study were drawn from the 926
assumption that impairments of tonic arousal, subjects described in Study I. Prior to each analysis,
a brief description of the relevant subsample is pro-
sensory registration, or the underlying percep- vided. Because the subjects either served as normal
tual abilities will also impede performance. controls in one of several studies, or were patients
For purposes of comparison, consider digit referred for clinical evaluations, the subsamples used
recall. Digit repetition involves a prototypic se- for each analysis differ (albeit with some overlap)
based on the specific neuropsychological measures
lective attention process within the verbal/ administered.
linguistic functional system. Like the BTA, out-
put is verbal rather than motoric. Until one's
immediate storage capacity is reached, repeating RESULTS
digits forward would appear to represent an un-
complicated selective attention test. Repeating Digit Span
digits backward, on the other hand, introduces a Pearson correlations between BTA scores and
requirement of simultaneous processing at the digit span (Wechsler, 1981) were based on a
end of each digit string presented. Thus, we pre- subsample of 452 patients and 149 normal sub-
dicted that BTA performance would correlate
86 DAVID SCHRETLEN ET AL.

jects, of whom 314 (52.2%) were male. The sub- patients and 44 normal control subjects, of
jects ranged from 6 to 86 years of age (M =38.1, whom 53% were men. Race was not recorded
SD = 19.7). Of the 572 subjects for whom race for one subject, but the remaining sample in-
was recorded, 59% were White, 40% were cluded 117 (52%) White subjects, 105 (47%)
Black, and 1% were "Other." The subjects had African-Americans, and 4 (2%) "Others." Sub-
completed a mean of 11.0, SD = 4.0 years of jects ranged in age from 15 to 72 (M = 37.9,Sp
schooling. The actual number of digits recalled, = 11.7) years. They had an average of 12.1, SD
both forward and backward, was used to com- = 3.3 years of education. As expected, BTA
pute correlations. As predicted, BTA scores cor- scores correlated significantly with all three
related more highly with backward (r = .53, p < parts of the Stroop (word reading, r = .66; color
.001) than forward (r = .43, p < .001) digit naming, r = .68; color-word naming, r = 67; all
span. 1A test for the difference between depen- ps < .001). However, BTA scores did not corre-
dent correlations (Bruning & Kintz, 1987) re- late more strongly with the interference trial
vealed that the disparity between these two cor- than with word reading or color naming trials.3
relations is statistically significant U(59?)= 3.12,
p < .01). Factor Analysis
To examine further the construct-related validity
Trail Making Test of the BTA, we conducted a principal compo-
Pearson correlations between BTA scores and nents analysis based on a subsample of 107 psy-
the Trail Making Test (Reitan, 1958) were based chiatric patients who completed a battery of neu-
on a subsample of 311 patients and 73 normal ropsychological tests. The patients ranged from
adults, of whom 59% were men. Of the 348 sub- 19 to 68 years of age, and included 62 (58%)
jects for whom race was recorded, 83% were men. Ninety patients (84%) were African-Amer-
White, 16% were African-American, and 1% ican, 13 (12%) were White, and the remaining 4
were "Other." Their ages ranged from 15 to 86 (4%) belonged to other races. They completed a
(41.1, SD = 15.4) years. As compared to the mean of 10.7, SD =2.6 years of school.
group used to examine correlations with digit All 107 patients completed a neuropsycholog-
span, this subsample contained more normal ical test battery from which 13 (raw scores) test
subjects, fewer patients with severe mental dis- variables were selected for entry into principal
orders, and fewer African-American subjects. components analysis. The scores of six WAIS-R
As predicted, total BTA scores correlated more subtests were chosen a priori and combined to
highly with Part B (r = -.55, p < .001)than Part create three composite scores: Verbal (Informa-
A (r= -.48,p = .001) of the Trail Making Test.2 tion + Similarities), Perceptual (Picture Comple-
Again, the difference between these correlations tion + Block Design), and Attention (Digit Span
was statistically significant (t(381)= 2.42, P < + Digit Symbol). In addition to BTA total sco-
.02). res, nine other cognitive test scores were entered
as variables in principal components analysis.
Stroop Color-Word Test These were Logical Memory and Visual Repro-
Correlations between BTA scores and the Stroop duction (immediate recall) from the Wechsler
Color-Word Test (Golden, 1978) were based on Memory Scale-Revised (Wechsler, 1987), a 30-
a subsample of 227 adults, which included 183 item short form of the Boston Naming Test
(Kaplan, Goodglass, & Weintraub, 1978), total
words recalled on the Hopkins Verbal Learning
1 When the results of nonnal subjects and patients
were analyzed separately, BTA scores consistently
correlated more highly with digits backward than dig- 3 When the results of patients were analyzed sepa-
its forward. rately, BTA scores correlated more highly with per-
2 When the results of normal subjects and patients formance on the interference trial than with perfor-
were analyzed separately, BTA scores consistently mances on the word reading and color naming trials.
correlated more highly with Trails B than Trails A. This was not the case for normal subjects.
 THE BRIEF TEST OF ATTENTION 87

Test (Brandt, 1991), the Rey-Osterrieth complex ferent functional neuropsychological systems.
figure drawing (Osterrieth, 1944), average time Thus, the BTA may tap an aspect of attention
to complete two trials of the Grooved Pegboard that is modality-independent in operation. This
test (K10ve, 1963) with the dominant hand, and would appear to involve the shifting of
all three trials of the Stroop Color-Word Test attentional resources back and forth between
(Golden, 1978). Principal components analysis two simultaneous but relatively simple cognitive
with varimax rotation yielded three factors with processes.
eigenvalues greater than unity (7.26, 1.08, and
1.01, respectively). This three-factor solution
accounted for 71.9% of the total variance and is CONCLUSIONS
highly interpretable. As shown in Table 2, Fac-
tor 1 accounted for the largest proportion of As shown in Study 1, the Brief Test of Attention
variance. This factor is aptly described as re- and its two constituent forms have acceptable
flecting general and verbal mental abilities. internal consistency based on the results of 829
Factor 2 clearly reflects attentional abilities, as patients and normal control subjects. Despite the
the BTA, WAIS-R Digit Span, and Digit Sym- modest correlations (.69 to .81) between Forms
bol composite scores, and all three Stroop scores Land N, group mean scores differed by less
loaded most highly on it. Factor 3 appears to be than .15 points in every sample examined, indi-
defined by perceptual abilities and psychomotor cating that the two forms are equally difficult.
speed. Further, individual between-forms difference
scores were normally distributed with a central
tendency of zero, suggesting that neither prac-
DISCUSSION tice nor interference effects influenced perfor-
mance from the first to second form adminis-
The results of this study suggest that the Brief tered. Test-retest reliability estimates are not yet
Test of Attention does, in fact, measure atten- available, but these data are being collected. The
tion. The task demands of the BTA are consis- distribution of scores produced by normal sub-
tent with the description of auditory divided at- jects, as shown in Table 1, must be regarded as
tention offered by Cooley and Morris (1990). provisional because the sample does not repre-
The BTA was found to correlate more strongly sent the U.S. population as a whole.
with the complex than simple components of Study 2 provides initial validation the BTA as
two accepted tests of attention that involve dif- a measure of auditory divided attention. Like

Table 3. Variable Loadings on Each Factor Derived from Principal Components Analysis.

Variable Factor I Factor 2 Factor 3

WA1S-R Verbal (lnf. + Sim.) .701 .30 .35

WA1S-R Attention (DSp. + DSym.) .47 .60 .37
WA1S-R Perceptual (PC + BD) .51 .15 .74
Boston Naming Test .68 .22 .25
Hopkins Verbal Learning Test .74 .22 .26
WMS-R Logical Memory .78 .32 .12
WMS-R Visual Reproduction .46 .13 .77
Brief Test of Attention .46 .61 .22
Stroop A (word reading) .36 .81 .09
Stroop B (color naming) .23 .84 .27
Stroop C (color-word interference) .15 .61 .50
Grooved Pegboard (dom. hand) .01 -.45 -.71
Rey-Osterrieth (copy) .35 .25 .71

1 The highest factor loading for each measure is shown in bold.

88 DAVID SCHRETLEN ET AL.

most neuropsychological tests, the BTA mea- lins Publishers.

sures general mental ability, as shown by its Cooley, E. L., & Morris, R. D. (1990). Attention in
children: A neuropsychologically based model for
substantial loading (.46) on the first factor that
assessment. Developmental Neuropsychology, 6,
emerged from principal components analysis. 239-274.
However, like all three parts of the Stroop test Cummings, J. L., & Benson, D. F. (1992). Dementia:
and the sum of WAIS-R Digit Span and Digit A clinical approach (2nd ed.). Boston: Butters-
worth.
Symbol, the BTA loaded most highly (.61) on a
Fleisher, L. S., Soodak, L. C., & Jelin, M. A. (1984).
factor that clearly is defined by attentional abili- Selective attention deficits in learning disabled
ties. Further evidence of the construct and crite- children: Analysis of the data base. Exceptional
rion validity of the BTA is presented in a com- Children, 51, 136-141.
panion article (Schretlen, Brandt, & Bobholz, Gibson, E. J. (1969). Principles ofperceptuallearn-
1996). ing and development. New York: Appleton-Cen-
tury-Crofts.
When applied to patients who are demograph- Golden, C. J. (1978). Stroop Color and Word Test.
ically similar to the normal subjects used in this Wood Dale, IL: Stoelting Co.
study, the Brief Test of Attention appears to pro- Hartlage, S., Alloy, L. B., Vasquez, c., & Dykman, B.
vide a reliable and valid appraisal of auditory (1993). Automatic and effortful processing in de-
divided attention. The entire test requires less pression. Psychological Bulletin, 113, 247-278.
Kaplan, E. F., Goodglass, H., & Weintraub, S. (1978).
than 10 minutes to administer and score, and
The Boston Naming Test. Boston: E. Kaplan & H.
when patients score < 4/1 0 or >7/1 0 on the first Goodg1ass.
form given, the second form may be omitted Klpve, H. (1963). Clinical neuropsychology. In F. M.
without compromising the reliability of clinical Forster (Ed.), The medical clinics of North Amer-
ica. New York: Saunders.
judgements regarding the presence or absence of
Mesulam, M.M. (1985). Attention, confusional states,
impaired attention. Until more broadly represen- and neglect. In M-M Mesulam (Ed.), Principles of
tative normative data become available, this re- behavioral neurology (pp. 125-168). Philadelphia:
port is intended to provide preliminary psycho- F. A. Davis Company.
metric information that will enable others to use Mirsky, A. F. (1989). The neuropsychology of atten-
tion: Elements of a complex behavior. In E.
the BTA for research and, with appropriate cau-
Perecman (Ed.), Integrating theory and practice in
tion, for clinical applications. clinical neuropsychology (pp. 75-91). Hillsdale,
NJ: Lawrence Erlbaum Associates.
Oscar-Berman, M., & Bonner, R. T. (1985). Match-
REFERENCES ing- and delayed matching-to-sample performance
as measures of visual processing, selective atten-
tion, and memory in aging and alcoholic individu-
Atkinson, R. C., & Shiffrin, R. M. (1968). Human
als. Neuropsychologia, 23, 639-651.
memory: A proposed system and its control pro- Osterrieth, P. A. (1944). Le test de copie d'une figure
cesses. In K. W. Spence & J. T. Spence (Eds.), The complexe. Archives de Psychologie, 30, 206-356.
psychology of learning & motivation: Advances in Posner, M. 1., & Snyder, C. R. R. (1975). Attention
research & theory (Vol. 2, pp. 89-195). New York:
Academic Press. and cognitive control. In R. L. Solso (Ed.), Infor-
mation processing and cognition (pp. 55-85).
Barkley, R. A., Grodzinsky, G., & DuPaul, G. J. Hillsdale, NJ: Erlbaum.
(1992). Frontal lobe functions in attention deficit Reitan, R. M. (1958). Validity of the Trail Making
disorder with and without hyperactivity: A review Test as an indication of organic brain damage. Per-
and research report. Journal of Abnormal Child ceptual and Motor Skills, 8, 271-276.
Psychology, 20, 163-188. Schretlen, D. (1989). The Brief Test of Attention. Bal-
Braff, D. L. (1993). Information processing and atten- timore, MD: Author.
tion dysfunctions in schizophrenia. Schizophrenia Schretlen, D., Brandt, J., & Bobholz, J.H. (1996). Val-
Bulletin, 19, 233-259.
idation of the Brief Test of Attention in patients
Brandt, J. (1991). The Hopkins Verbal Learning Test: with Huntington's disease and amnesia. The Clini-
Development of a new memory test with six equiv- cal Neuropsychologist. This issue.
alent forms. The Clinical Neuropsychologist, 5, Stroop, J. R. (1935). Studies of interference in serial
125-142.
verbal reactions. Journal of Experimental Psychol-
Bruning, J. L., & Kintz, B. L. (1987). Computational
ogy, 18, 643-662.
handbook of statistics. Glenview, IL: Harper Col- Stuss, D. T., & Gow, C. A. (1992). "Frontal dysfunc-
 THE BRIEF TEST OF ATTENTION 89

tion" after traumatic brain injury. Neuropsychia- logical Corporation.

try, Neuropsychology, and Behavioral Neurology, Wechsler, D. (1987). WechslerMemory Scale-Revised
5,272-282. manual. San Antonio: The Psychological Corpora-
Triesman, A. M. (1964). Selective attention in man. tion.
British Medical Bulletin, 20, 12-16. Wickens, C. D. (1984) Processing resources in atten-
Wechsler, D. (1939). Measurement of adult intelli- tion. In R. Parasuraman & D. R. Davies (Eds.),
gence. Baltimore: Williams & Wilkins. Varieties of attention (pp. 63-102). Orlando, FL:
Wechsler, D. (1981). The Wechsler Adult Intelligence Academic.
Scale-Revised manual. San Antonio: The Psycho-

View publication stats