Perceptulrl and Motor Skills, 1971, 3 3 , 575-581.

@ Perceptual and Motor Skills 1971

TRAIL MAKING TEST RESULTS FOR NORMAL AND

BRAIN-DAMAGED CHILDREN1
RALPH M. REITAN'
University oof Warhingron
Summary.-The Trail Making Test ( a shortened version for use with children) was administered to 98 normal children who ranged in age from 109 to
179 mo. N o significant differences between the performances of boys and girls
were found. Relatively weak but significant relationships between scores on
both Parts A and B and chronological age were present. The second parc of the
srudy compared 35 matched pairs of Ss with and without cerebral lesions. The
non-brain-damaged Ss had scores comparable to those of normal children tested
in the first part of the srudy, but the brain-damaged children performed much
more poorly. The results indicate that the skills required by either Part A or
Part B of the Trail Making Test are significantly impaired by cerebral damage.

The Trail Making Test was originally published by the Adjutant General's
Office, War Department, U. S. Army, in 1944 as one of the Performance subtests of the Army Individual Test. Armitage ( 1946) included this test among
others for evaluating the effects of brain injury among soldiers. He found distinctly positive results and a number of other investigations have since been performed to determine the extent to which results on this test differentiate adult
Ss with and without brain lesions. Reitan published two general studies of this
type (1955, 1958). Other aspects of brain-damage have also been explored,
including the effects of acuteness vs chronicity of cerebral damage (Fitzhugh,
Fitzhugh, & Reitan, 1962), the significance of dysphasia as compared with absence of specific language deficits in brain-damaged Ss (Reitan, 1960), and the
differential influence of right vs left cerebral lesions (Reitan & Tarshes, 1959;
Fitzhugh, Fitzhugh, & Reitan, 1963; Wheeler, Burke, & Reitan, 1963; Wheeler
& Reitan, 1963). Other conditions and disorders, associated with psychological
deficit, have also been explored. Several studies have investigated Trail Making Test results as a function of adult aging (Reitan, 1962; R e e d & Reitan,
1963a; Reed & Reitan, 1963b; Fitzhugh, Fitzhugh, & Reitan, 1964). Ss with
chronic alcoholism have been compared with brain-damaged Ss by Fitzhugh, Fitzhigh, and Reitan ( 1960), whereas Alvarez has performed a similar type of comparison between depressed and brain-injured Ss (1962). The above studies have
generally yielded positive results in terms of the efficacy of the Trail Making
Test with respect to reflecting psychological deficit in persons with brain damage and the other conditions mentioned. Conversely, Brown, et al. (1958) reported a failure to find differences between groups of Ss with cerebral lesions
and schizophrenia.
'This study was supported in parc by Grants NS 09897 (National Institute of Neurological Diseases and Blindness) and HD 02274 to the Child Development and Mental Retardation Center (National Institute of Child Health and Human Develooment).
T h e author is indebted to Miss Jan Janesheski for assistance with statistical computations.

5 76

R. M. REITAN

A version of the Trail Making Test for use with children was prepared by
Reitan in 1955 merely by shortening the adult form. Davids, Goldenberg, and
Laufer (1957) have studied this form of the test, with findings which indicated
that brain-damaged children performed more poorly than normal children. The
present study was undertaken because the results on adults appear promising in
general and the above study by Davids, et al. is the only one on children presently available in the literature.
The first part of the investigation was based on normal children whereas
the second part compared children with and without brain damage. Among
the normal children, boys and girls were compared with regard to their performances on both parts of the test and correlations were done between chronological age and both parts of the test for boys and girls. Analysis of results obtained
with normal children provided a background for evaluation of the differences
between children with and without cerebral damage.

METHOD

Trail Making Test

Both Parts A and B of the Trail Making Test consist of a single sheet of
paper with a sample printed on the front side and the test itself on the back side
of the sheet. In Part A 15 circles are distributed over the page. Each circle encloses a number and these range from 1 to 15. The adult version of Part A
uses 25 circles and the children's version was constructed merely by duplicating
only the first 15 circles. S's task is to connect the circles in numerical sequence
as quickly as possible with a penciled line. The score is the number of seconds
required for completion of the task. S is ca~~tioned
to be careful to avoid
mistakes. However, if he makes a mistake the examiner calls it to his attention
immediately, indicates the last circle correctly reached, and asks S to proceed from
that point. Thus, in practice, mistakes contribute to the time required for completion of the task. In P u t B of the Trail Making Test there ate also 15 circles
on a sheet' of paper, but these circles contain both numbers and letters. S is
directed to begin at number 1, to draw a line to A, then to 2, next to B, and so
on, alternating between numbers and letters until he reaches the end. The revision from the adult version again consists merely of having used only the first
15 circles of the task. The score is the amount of time required to complete the
task and errors are handled in the same way as on Part A.

Sgbjects
The normal Ss were 51 boys and 47 girls who ranged in age from 109 mo.
to 179 mo. These were voluntary Ss who responded to a request sent to patents
for normal Ss to participate in a psychological testing study. They were not
compensated and were told in advance that they would not be informed of the
results. Only 2 boys and 2 girls were 9 yr. of age, so the essential distribution
for both boys and girls ranged from the 10th birthday through the 14th birthday.

TRAIL MAKING BY CHILDREN

577

The mean age for boys was 151.65 mo. and for girls was 151.02 mo., with standard deviations of 17.41 and 19.51, respectively. Comparison of the difference
between these means yielded a t ratio of 0.17 ( p < .go). Educational levels
were computed on the basis of the last grade completed. The boys had a mean
education of 6.07 yr. and the girls 6.04 yr. (SDs, 1.27 and 1.54, respectively).
Comparison of the difference between these means yielded a t ratio of 1.07 ( P
< .30). Thus, the boys and girls were essentially similar in age and education.
For the second part of this study, in which groups of Ss with and without
cerebral lesions were compared, 35 Ss with known brain damage were matched
for color, sex, and age with control Ss for whom there was neither past nor present evidence of neurological damage or disease. The brain-damaged group had
a mean age of 141.26 mo. and the control group had a mean age of 141.23 mo.
(SDs of 20.65 and 18.39, respectively). Since education might well be considered a dependent variable with respect to brain damage, no attempt was made
to match the groups for educational achievement. The age difference between
the groups did not approach statistical significance. Ne~uologicaldiagnoses for
the 35 brain-damaged Ss were distributed as follows: traumatic injury, 12; tumor,
7; congenital vascular anomaly, 5;'inflammatory damage, 5; perinatal lesions, 5;
and cerebral abscess, 1. The five instances of perinatal lesions were diverse and
included birth injury, spastic right hemiplegia, right frontal-temporal fibrous
dysplasia, and two instances of demonstrated right cerebral atrophy.
Ss were tested by psychological technicians who had no advance knowledge
of the specific research uses for which the results were intended. This procedure may have guarded against unintended influencing of the results in accordance with the research hypotheses during the data collection phase.

Procedure
The procedure involved evaluation and analysis of Trail Making Test results for normal boys and girls, followed by a comparison of the results obtained
from Ss with and without cerebral damage. Means and standard deviations were
computed for Parts A and B for all groups tested. Among the normal Ss means
on both parts were compared for girls and boys. In addition, Pearson productmoment coefficients of correlation for boys and girls were computed between
chronological age and Parts A and B. Further, comparisons of possible significant differences between the coefficients of correlation for boys and girls were
made by converting the coefficients to Fisher's z values, determining the standard
error of z,and computing t ratios.

RESULTS
The means and standard deviations for normal boys and girls, together with
t ratios comparing these groups, are presented in Table 1. Mean values expressed
in terms of seconds required for completion of the cask were almost identical
for boys and girls on both Parts A and B. Thus, it appears that the present

TABLE 1
MEANS, STANDARDDEVIATIONS,
AND t RATIOSFOR NORMAL
BOYS AND
ON PARTSA AND B OF THE TRAILMAKINGTEST

Part A

Boys

Girls

<.90

0.54

<.60

SD

12.05

14.85
6.82
31.51
14.57

0.1G

15.04
4.76
32.98

M*

SD
Part B

GIRLS

*Seconds.

PART A

- Boys

--

r = -.33 (p<.05 1
Girls r = -.43 (pC.01 1

a--ooaC(

-- Girls r = -.60(

11

12

10

13

14

15

D <.01)

C h r o n o l o g i c a l age
FIG. 1. Relationships between Trail Making Test performances and chronological
age for normal boys and girls

TRAIL MAKING BY CHILDREN

5 79

study generated little evidence for a sex-related difference on this task within
the age range used.
The relationship of Trail Making Test performance and chronological age
is shown in Fig. 1. The first part of the figure is based on data from Part A and
the second part of the figure from Part B of the test. It is apparent that both
boys and girls show a tendency to be able to perform the test more quickly as
they grow older. With boys, however, the coefficients for both Parts A and B
were lower than for @;ls. Comparison of the coefficients for Part A yielded a
t ratio of 0.56 ( p < .GO) whereas for Part B the t ratio was 1.78 ( p < .lo).
Thus, correlations for boys' and girls' scores were not significantly different, although in each comparison the magnitude of the coefficient was greater for girls
than boys.
TABLE 2

MEANS,STANDARD
DEVIATIONS,
AND t RATIOS FOR MATCHED GROUPS
WITH AND WITHOUT
CEREBRAL
LESIONS ON PARTSA AND B
OF THE TRAILMAKINGTEST
Part A

SD
Part B

Brain-damaged Control
37.06
16.74
29.79
110.19
85.91

8.12
35.03
15.76

3.75* <.001

5.15t <.OOl

4.48*

5.06t <.001

<.001

SD
*Computations based on raw-score distributions (sec.).
tComputations based on normalized T-score distributions.
Table 2 presents means, standard deviations and comparisons of the groups
with and without cerebral damage. It is apparent that the control group performed almost as well as the boys and girls in the study of normal children.
However, the children with brain damage did considerably more poorly. Comparison of the normal and brain-damaged groups on Parcs A and B yielded significant t ratios. In order to explore the possibility that these t ratios were limited by non-normality of distributions and corresponding influences on the denominator of the significance ratio, the scores for the combined groups were
converted to normalized T-score distribution, following McCall's method which
yields a mean of 50 and standard deviation of 10 (McNemar, 1949). Comparisons of these distributions are also presented in Table 2. The t ratios based
on these normalized distributions are somewhat larger than those based on the
raw scores, but in both cases they were significant. Thus, it appears that both
Parts A and B of the Trail Making Test reliably reflect deficits in performance
in Ss with cerebral damage.

DISCUSSION
The Trail Making Test requires a degree of alertness in visual comprehension of stimulus material, interpretation of the symbolic significance of the stim-

580

R. M. REITAN

ulus material (numbers and letters), visual scanning in order to be able to locate
the next number or letter that S is seeking, and in Part B, the ability to keep in
mind a numerical series and an alphabetical series simultaneously. While Part B
has usually shown somewhat more striking differences than Part A in comparisons of groups of adult Ss with and without cerebral lesions (Reitan, 1955,
1958), no such difference appears in this study. The raw score distributions
yielded a somewhat larger t ratio for Part B than A, but probably no difference
in the groups is present with respect to their comparative performances as
judged by the comparable ratios derived from the normalized distributions. The
present results yield strong confirmation of the findings by Davids, et al. (1957),
indicating that this test is sensitive to brain damage in children.
The relationship among normal children between test results and chronological age suggests that this is the type of task that most 10-yr.-old children can
perform relatively well. The correlation was somewhat increased for girls on
Part B, again probably reflecting the fact that with the more difficult task
the performances had less tendency to have become asymptotic by the age of
10 yr. Consideration of the actual means for brain-damaged children, however,
suggests that their performances are considerably impaired even with respect
to mean performances of the younger normal children in this scudy. Thus, brain
damage, within the age range of this study, appears to be a more potent determiner of the test results than chronological age. It must also be noted, however,
that brain damage may be relatively mild in nature or relatively severe. It was
not possible with the present design to evaluate severity of brain damage as a
factor in relation to Trail Making Test performances, but it seems reasonable to
postulate that relatively mild cerebral damage might be associated with relatively
normal performances. This type of consideration, in turn, reflects the necessity
for stating that probably no single test represents an adequate behavioral framework for expression of brain functions (Reitan, 1966). In other words, the results of this study should in no sense be interpreted to imply that the Trail Making Test, by itself, is an adequate instrument for reflecting brain-behavior relationships in children.
~

REFERENCES
ALVAREZ,
R. R. Comparison of depressive and brain-injured subjects on the Trail Mak-

ing Test. Perceptual and Motor Skilh, 1962, 14,91-96.

AMITAGE? S: G. An analysis of certain psychological tests used for the evaluation of
b r a ~ nInjury. Psychological Monogrnaflhs, 1946, 60, No. 1 (Whole No. 277).
BROWN,,
E. C., CASEY, A., FISCH? R. I., & NEURINGER,
C. Trail Making Test as a screenlng devlce for the detecuon of b r a ~ ndamage. Journal of Consulting Psychology,
1958, 23,469-474.
DAVIDS,
A., GOLDENBERG,L., & LAUFER, M. W. The relation of the Archimedes spiral
after-effect and the Trail Making Test to brain damage in children. Journal of
Consulting Psychology, 1957, 21,429-433.
FITZHUGH,
K. B., ,FI?ZHUGH,,
L. C., & ,REITAN,R. M. The relation of acuteness of
organic b r a ~ ndysfunct~onto Trall Mak~ngTest performances. Perceptrdal and
Motor Skills, 1962, 15, 399-403.

TRAIL MAKING BY CHILDREN

581

FITZHUGH,K. B., FITZHUGH,L. C., & REITAN,R. M. Effects of "chronic" and "current" lateralized and non-lateralized cerebral lesions upon Trail Making Test performances. Journal o f Nervous and Mental Disease, 1963, 137, 82-87.
FITWUGH, K. B., FITZHUGH,L. C., & REITAN,R. M. Influence of age upon measures
of problem solving and experiential background in subjects with longstanding
cerebral dysfunction. ]ournu1 o f Gerontology, 1964, 19, 132-134.
FITZHUGH,L. C., FITZHUGH,K. B., & REITAN, R. M. Adaptive abilities and intellectual functioning in hospitalized alcoholics. Quarterly ]ournu1 o f Studies on
Alcohol, 1960, 21, 414-423.
MCNEMAR, Q. Psychological statistics. New York: Wiley, 1949.
REED, H. B. C., & REITAN, R. M. A comparison of the effects of the normal aging
process with the effects of organic brain damage on adaptive abilities. Journal o f
Gerontology, 1963, 18, 177-179. ( a )
REED, H. B. C., & REITAN, R. M. Changes in psychological test performances associated with the normal aging process. Journal o f Gerontology, 1963, 18, 271-274.
(b)
REITAN, R. M. The relation of the Trail Making Test to organic brain damage. Journal o f Consulting Psychology, 1955, 19, 393-394.
REITAN, R. M. The validity of the Trail Making Test as an indicator of organic brain
damage. Perceptual and Motor Skills, 1958, 8, 27 1-276.
REITAN, R. M. The significance of dysphasia for intelligence and adaptive abilities.
Journal o f Psychology, 1960, 50, 355-376.
REITAN, R . M. The comparative psychological significance of aging in groups with and
without organic brain damage. In C. Tibbetts & W . Donahue (Eds.), Social and
prychological aspects o f aging. New York: Columbia Univer. Press, 1962. Pp.
880-887.
REITAN, R M. Problems and prospects in studying the psychological correlates of brain
lesions. Cortex, 1966, 2, 127-154.
REITAN, R. M., & TARSHES,E. L. Differential effects of lateralized brain lesions on
rhe Trail Making Test. Journal of Nervous and Mentul Disease, 1959, 129, 257262.
WHEELER,L., BURKE,C. J., & REITAN,R. M. An application of discriminant functions
to the problem of predicting brain damage using behavioral variables. Perceptual
and Motor Skills, 1963, 16, 417-440. (Monograph Supplement 3-V16)
WHEELER,L., & REITAN,R. M. Discriminant functions applied to the problem of predicting cerebral damage from behavioral tests: a cross-validation study. Perceptual
and Motor Skills, 1963, 16, 681-701.
Accepted June 7,1971