Large Sample Population Age Norms for Visual Acuities
Obtained With Vistech-Teller Acuity Cards
Solange Rios Salomao*^
and Dora Fix VenturafX
Purpose. To determine population age norms in the first three years of life for binocular and
monocular grating visual acuity (VA) obtained with Vistech-Teller Acuity Cards (TAC).
Methods. TAC was used to estimate grating acuity in 646 healthy infants and children born
at due date ± 2 weeks, all of whom underwent ophthalmologic and orthoptic evaluation. The
sample consisted of 20 age groups from 0 to 36 months. Sixty-nine percent of the children
attended day care centers in the city of Sao Paulo. The sample was composed of white (63.0%),
mulatto (25.2%), African-Brazilian (11.0%), and Asian (0.8%) infants and children, most of
whom (97%) were from low-income families. Tests were conducted by eight highly trained
testers, six of whom were orthoptists.
Results. Binocular and monocular norms for grating VA are presented in terms of tolerance
limits for 90% of the population with 95% probability. The range of tolerance limits is
approximately 2.5 octaves at most ages. There were no statistical differences among scores
obtained by the different testers. There were no differences in VA due to race, sex, and first
or second eye tested. The results on binocular (99.3%) and monocular (96.2%) testability
and on mean test duration (13 minutes for one binocular and two monocular measurements)
confirm the clinical applicability of TAC.
Conclusions. The binocular and monocular grating VA norms obtained in this large-sample
study are different from the preliminary norms published with the TAC. Results from this
and other studies (see Mayer et al, page 671, this issue) strongly point to a need for redefinition
of the preliminary VA norms. Invest Ophthalmol Vis Sci. 1995; 36:657-670.
Uehaviorally measured grating visual acuity (VA),
which reflects the ability to resolve the elements of a
pattern, develops rapidly from birth to 6 months and
then at a much slower rate up to adult acuity at 3 to 5
years of age.1"4 There is no question about the clinical
importance of assessing visual acuity at an early age.
Early detection and proper correction of visual abnormalities, such as congenital cataract and strabismus,
can prevent permanent visual impairment.5"9 Reversibility of impairment has been shown to occur in ani-
From the * Department of Ophthalmology, Paulista School of Medicine, Hospital SOo
Paulo; the fCenter for Neuroscience and Behavior, University of Sdo Paulo; and the
XDepartmenl of Experimental Psychology, Institute of Psychology, University of Sao
Paulo, Sao Paulo, Brazil.
Presented in part at the annual meeting of the Association for Research m Vision
and Ophtlialmology, Sarasota, Florida, 1993.
Supported by Fundacao de Amparo a Pesquisa do Estado de Sdo Paulo grant 88/
3224-3.
Submitted for publication February 10, 1994; revised September 28, 1994; accepted
October 21, 1994.
Proprietary interest category: N.
Reprint requests: Dora Fix Ventura, Instituto de Psicologia, Universidade de Sao
Paulo, A-u. Prof. Mello Moraes 1721, Sdo Paulo, SP 05508-900, Brazil.
Investigative Ophthalmology & Visual Science, March 1995, Vol. 36, No. 3
Copyright © Association for Research in Vision and Ophthalmology
mal models of early abnormal visual experience,10 and
clinical application of these findings has been confirmed.11
Behavioral grating visual acuity can be measured
successfully in infants in the laboratory using a variety
of psychophysical procedures' 2 " 17 based on pattern
preferences first described by Fantz18 in 1958. Measurement of grating acuity in the clinical setting required the development of a rapid measuring technique. The acuity card procedure (ACP) developed
in 1985 by McDonald and colleagues'9"21 provided
a potential answer to this problem. It involved the
construction of commercially available equipment
(Vistech-Teller Acuity Cards [TAC]) and used an informal but rapid testing procedure. Preliminary binocular norms presented with the test were compiled
from data from six different studies19"24 with a total
sample of 249 children. Preliminary monocular norms
were based on data from two studies from the same
laboratory,20>2! with a sample of 72 children. Acuity
norms were presented at 1- to 6-month intervals.25
657
658
Investigative Ophthalmology & Visual Science, March 1995, Vol. 36, No. 3
Since its release, ACP has been used extensively
both in research and in clinical practice. Data based
on ACP are already available on various methodologic
aspects of its application (use in private practice, testretest reliability, intertester agreement, special procedures for use with visually impaired children),26"28 on
normal visual acuity development,29"31 on comparison
with refractive errors,22'32"34 on application to different ocular pathologies,28'35"40 and on assessment of
visual acuity in children with neurologic impairment.4142 This large applicability generated the need
for a review of the original preliminary norms widely
used in clinical practice and clinical research at present and the need for large scale studies29"31 specifically
designed for the purpose of establishing more statistically valid age norms.
To establish normative data, two conditions are
basic. The first consists of ruling out the possibility of
including pathologic cases in the sample of normal
children. This can be accomplished by submitting every tested child to ophthalmologic and orthoptic evaluation, a requirement not fulfilled by several studies
that aimed at defining norms.29"31 The second is to
have a sufficiently large sample that can be representative of the population. The present study met these
conditions, and its purpose was twofold: to establish
precise tolerance limits for grating acuity measured
with TAC in healthy children from 0 to 36 months
and to define a procedure in which testability, testing
time, and use of a relatively large number of testers
were compatible with the use of TAC for screening
and clinical purposes.
METHODS
Subjects
The final sample consisted of 646 children, all of
whom contributed data to the binocular norms and
624 who contributed data to the monocular norms.
They ranged from 0 to 36 months of age, with nominal
age and sample size per age group, as in Tables 1 and
2. Age range was ±1 month for all age groups above
6 months of age. The children were tested in three
different places: the Department of Ophthalmology
at the Paulista School of Medicine-Hospital Sao Paulo
(N = 123; 19%), the Department of Experimental
Psychology at University of Sao Paulo (N = 75; 12%),
and nine public day care centers in the city of Sao
Paulo (N = 448; 69%).
Racial categories were: white (N = 407, 63.0%),
mulatto (N = 163, 25.2%), African-Brazilian (N =
71, 11.0%), and Asian (N = 5, 0.8%). The present
sample was classified according to and follows the distribution of the racial groups that exist in the Brazilian
population according to the last census published in
1983.43 Most (N = 626, 97%) were from low-income
and the remaining (N = 20, 3%) were middle-to-high
income families.
Subjects had no known neurologic or medical abnormalities and had no previous ophthalmologic assessment. Complete ophthalmologic examination, including external eye inspection, cycloplegic retinoscopy, indirect ophthalmoscopy, and corneal reflexes,
was conducted. All children were also submitted to
orthoptic evaluation, which consisted of assessing ocular alignment and motility.
The final sample was composed of infants and
children that met the following criteria:
1. Birth date coincided with due date or was ±2
weeks from it. In the latter case, the age was
corrected.
2. Emmetropia or refractive error was within < 5
spherical diopters of hyperopia, < 3 spherical
diopters of myopia, < 3 cylindrical diopters of
astigmatism, and < 2 diopters of anisometropia
(spherical equivalents).
3. There was an absence of fundus alterations, apparent ocular deviation, and ocular diseases.
An additional 80 children were excluded from the
final sample either because they did not undergo the
eye examination (N = 47), they did not meet our
criteria on the eye examination (N = 22), they were
not born within 2 weeks of due date (N = 9), or
because they did not complete the binocular test of
acuity (N = 2).
The tenets of the Declaration of Helsinki were
followed in the present study. For children who were
brought individually to the testing place, the procedures were explained and informed consent was obtained from each subject's parent(s) or guardian(s)
before testing. In public day care centers, a meeting
with directors, nursemaids, and parents or guardians
was arranged to explain the reasons for early visual
assessment and the type of procedures to which the
child would be submitted. Only children whose parents or guardians attended the meeting were tested.
Three parents or guardians did not agree with the
procedures, and they were asked to leave a written
statement.
Apparatus and Procedure
Grating acuity was assessed with a variant of the acuity
card procedure binocularly and monocularly. The
first measurement was always binocular. The first eye
tested monocularly was randomly selected (OD was
the first eye tested in 338 cases -54.2%). Monocular
measurements were made with an adhesive eye patch
(OFTAM, AMP, Sao Paulo, Brazil). Infants up to 20
days of age were tested only binocularly. Details of the
•r-
Age Norms for Vistech-Teller Acuity Cards
659
TABLE l. Binocular Norms for Vistech/Teller Acuity Cards
Age
(months)
0.5
2
3
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
Total
N
17
9
21
25
30
34
35
40
34
41
35
37
36
44
37
42
35
25
37
32
646
Test
Distance
Start Card
(cyc/deg)
Mean Acuity
(cyc/deg)
Lower Limit*
(cyc/deg)
Closest Card
(cyc/deg)
38
38
38
38
38
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
0.30
0.30
0.44
0.44
0.44
1.3
1.3
1.3
1.3
1.3
1.3
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
0.66
2.02
3.89
5.48
7.44
9.81
11.59
11.08
13.04
13.08
12.39
13.81
14.76
14.64
16.66
15.28
17.24
17.36
19.19
17.82
0.27
0.23
1.59
2.14
3.38
5.10
6.05
4.13
5.63
5.91
5.10
6.60
7.68
6.20
8.92
7.25
8.61
9.96
9.50
7.12
0.22
0.22
1.1
1.7
3.1
3.1
4.7
3.1
4.7
4.7
4.7
4.7
6.4
4.7
6.4
6.4
6.4
6.4
6.4
6.4
* In the finit three age groups, tolerance limits are calculated for 90i% of the population with 9 0% probability.
acuity card apparatus have been described elsewhere.44
The subject was seated in front of a gray screen
(Vistech, Dayton, OH). A rectangular opening in the
middle of the screen was used to present the acuity
cards. Children 0 to 6 months of age were positioned
38 cm away from the screen, and children from 7 to
36 months of age were positioned 55 cm from the
screen. The luminance of the cards was measured with
a photometer (Sekonic Auto Lumi L-158) and was
greater than 10 cd/m 2 (grating acuity is not affected
by changes in luminance above this value40).
A standard set of 16 acuity cards (Vistech, Dayton,
OH) with half-octave steps (0.23 to 38 cyc/cm) was
used. The cards were arranged face down on a table
behind the screen, with low spatial frequencies (wide
stripes) at the top of the stack and progressively higher
spatial frequencies below. The tester received each
card from an assistant who selected the start card
based on the age of the child according to the specification from the TAC Handbook44 (see Tables 1 and
2). The tester was not informed about the start card
value and was blind to the left-right position of the
grating on it.
The task of the tester was to judge from the subject's behavioral reactions—such as eye movements,
* Children were not tested at distances beyond 55 cm, as is
sometimes done to avoid a "ceiling" effect at the 38 cyc/cm card,
because there were very few thresholds at that card. They were seen
only in the last four age groups (total, 128 children) and correspond
to 6.3% of binocular and 3.1% of monocular visual acuity tests.
head movements, and pointing—which side of the
card displayed the grating. Three responses were available: left, right, and no response (child cannot resolve
grating). Cards were presented by the tester with several reversals of left-right position as many times as
the child found necessary for a judgment. A correct
judgment was considered a positive response; an incorrect judgment or the inability to resolve the grating
was considered a negative one. The assistant provided
the tester with feedback as to the correctness of the
judgment and moved on to the next card.
A modified staircase procedure was used to determine the grating visual acuity threshold. Cards were
presented from lower to higher spatial frequencies in
1-octave steps up to the threshold region, and then at
0.5-octave steps around the threshold. After a negative
result (tester reported "not seen" or made a wrong
judgment), step size was reduced from 1- to 0.5-octave,
and a lower spatial frequency was presented until a
positive response was obtained. Testing was continued
until two consecutive staircase reversals occurred. VA
threshold was defined as the spatial frequency of the
last card that received two positive responses. If this
criterion was not met in six trials (counting from the
first negative response), threshold was defined as the
highest spatial frequency card that received the greatest number of positive responses. The tester classified
confidence of each grating acuity assessment according to a 5-point subjective scale, ranging from 1
(very low) to 5 (very high).
A test consisted of one binocular followed by two
660
Investigative Ophthalmology & Visual Science, March 1995, Vol. 36, No. 3
TABLE 2. Monocular Norms for Vistech/Teller Acuity Cards
Age
(months)
2
3
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
Total
N
8
21
25
30
34
35
40
34
41
35
37
35
43
37
42
34
24
37
32
624
Test
Distance
Start Card
(cyc/deg)
Mean Acuity
(cyc/deg)
Lower Limit*
(cyc/deg)
38
38
38
38
55
55
55
55
55
55
55
55
55
55
55
55
55
55
55
0.44
0.44
0.44
0.44
1.3
1.3
1.3
1.3
1.3
1.3
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.31
3.09
4.15
7.18
8.31
10.88
9.82
10.35
10.07
9.95
11.12
12.09
12.31
13.00
12.79
14.12
15.22
14.97
14.98
0.42
1.18
1.57
2.78
4.31
4.26
4.00
4.82
4.70
5.21
3.91
4.30
4.03
6.62
5.61
7.09
7.99
7.46
5.86
Closest Card
(cyc/deg)
0.6
1.1
2.2
2.2
3.2
3.2
3.2
3.2
3.2
4.7
3.2
4.7
3.2
4.7
4.7
4.7
6.4
6.4
6.4
' III the first two age groups, tolerance limits are calculated for 90% of the population with 90% probability.
monocular grating acuity measurements. Total test
time was measured from the beginning of binocular
measurement to the removal of the occluder in the
end of the second monocular test. It should be noted
that the 2-week-old group would be expected to have
the lowest total test times because they were tested
only binocularly.
After grating acuity assessment of the subjects, the
orthoptic and ophthalmologic examinations were performed, with both professionals unaware of each other's results. The ophthalmologist was unaware of the
grating acuity results. The cycloplegic agents used
were cyclopentolate (1 drop/eye) plus tropicamide
0.5% (1 drop/eye).
Training of the Testers
A team of eight acuity card testers was formed, two of
whom had extensive experience with TAC. One of
them and an author of this article (DFV) gave theoretical and practical training to the other six testers—
four orthoptists and two psychology undergraduate
students. During training, each tester had to measure
binocular and monocular grating acuity in five normal
children and five children with different ocular disorders, ranging in age from 0 to 36 months. No age
group was privileged in this phase. To be included as
a tester, at the end of the training period each participant had to display errorless performance in the
"scrambled cards" procedure.44 The number of subjects/tester were: AB - N = 236 (36.5%); EG - N =
178 (27.5%); MM - N = 108 (16.7%); VF - N = 54
(8.3%); SS - N = 25 (3.9%); RE - N = 21 (3.3%);
FC - N = 21 (3.3%), and RO - N = 3 (0.5%).
Retest
A retest was performed when one or more of the following conditions occurred: the subject did not cooperate during the test; acuity scores were under the age
norm by more than 1 octave; there was an interocular
acuity difference (IAD) equal to or greater than 1
octave or a confidence of grating acuity estimate lower
than 3 octaves (fair). The retest was always performed
by a different tester who, though aware that it was a
retest, received no information about the previous test
score or the reason for the retest. It should be noted
that the norms were based on the test results of all
subjects from the final sample, regardless of the fact
that part of the subjects were retested.
Statistical Analysis
All data analysis was conducted using log values of the
VA scores. Norms were calculated based only on test
results. Retest data were not included.
Norms were described using tolerance limits,46
which assume normal distribution of the function in
the population. These are intervals that cover a previously fixed percentage of the population for the
function under study, with an attached probability.
For the present study, tolerance intervals for 90% of
the population with a probability of 95% were used.
The exception was for the 2-week-, 2-month-, and 3month-old groups, for whom a probability of 90% was
Age Norms for Vistech-Teller Acuity Cards
adopted because there is a steep rise in acuity in this
phase and the N was small in these groups. After adherence to the normal curve was tested and confirmed
for the group with the largest N, the determination
of tolerance limits was made using an N close to 38
subjects in each age group, as required for tolerance
limits with the specified probabilities (N for each
group is presented in Tables 1 and 2). Tolerance limits
were calculated with the equation: TL] = mean — K
X SD and TLU = mean + K X SD, where TL, and TLV,
are, respectively, the lower and upper tolerance limits;
K is a constant based on the value of N (number of
subjects), the desired probability, and the percentage
of the population covered; and SD is the standard
deviation of the mean.46
Analysis of variance (ANOVA) was performed using VA scores to evaluate agreement among testers,
including age as a factor, among VA scores from both
eyes obtained by the eight TAC testers separately for
monocular and binocular VAs; test IADs across the 19
age groups; evaluate possible differences among racial
categories, sex, and first and second eye tested in monocular VA; and compare differences between binocular and monocular VA in the 20 age groups. The Pearson coefficient was calculated to verify correlation between refractive errors of OD versus OS and refractive
errors versus VA in the entire sample. All analysis performed used test results. Retest data were used only
to compare with test data on the same children.
RESULTS
Population Norms
Tolerance Limits. Data from the final sample of 646
children were used to calculate binocular tolerance
limits. The monocular tolerance limits were based on
data from 624 of these children. Both limits were determined from VA scores taken from only one eye per
subject, randomly selected between first and second
eyes tested in each child. The rationale for randomly
selecting VA between eyes is that both eyes should not
be used because they are not independent,47 and using
only the first eye tested is also inappropriate because
practice and fatigue might affect variability.48 The first
eye tested was randomly selected so that left and right
eyes would be equally represented in the sample. Binocular and monocular VA means with respective SEs
and tolerance limits are shown for each age group in
Tables 1 and 2 in cyc/deg and in Figure 1.
A one-way ANOVA for binocular VA (Fi9645 =
89.72, P= 0.01) and monocular VA (F18,623 = 31.54; P
= 0.01), followed by the Scheffe procedure, identified
pairs of age groups that were significantly different at
the 0.05 level. For monocular and binocular VA, 2month-olds had significantly lower acuity than 4-
661
month-olds, who, in turn, had significantly lower acuity than 6-month-olds (P < .05 for all). After 6 months,
there was a gradual increase in acuity with no significant differences between adjacent age groups (P >
.05 for all).
Tables 1 and 2 present, respectively, binocular and
monocular norms and the closest card to the lower
limit in cyc/deg for clinical use. Tolerance limits are
also presented, along with the preliminary norms44 in
Figure 2 for binocular and monocular VA scores. A
steep increase in binocular and monocular VA is observed from birth to approximately 6 months, followed
by shallow growth thereafter. The present results exhibit much higher monocular VAs than the preliminary norms in the first 18 months of life. The lower
limit was higher by about 1 octave between 8 and 14
months and by less than 1 octave from 4 to 6 months.
As to the upper limit, the difference was close to 2
octaves from 2 to 24 months. From 28 to 34 months,
the lower limit was lower than the preliminary norms
by less than 1 octave, and for the last age group the
difference was slighdy higher than 1 octave.
Interocular Acuity Differences. Mean IADs (OD mi-
nus OS) for the 19 age groups are shown in Table 3,
along with SDs and tolerance limits for IADs. Absolute
IAD values are also shown. Differences ranged from
-0.15 octaves (OS better) to 0.16 octaves (OD better).
In absolute IAD values, the range was from 0 to 3
octaves. These differences were not restricted to specific age group(s). A one-way ANOVA based on the
raw IADs showed no significant differences among the
19 age groups (F18,623 = 0.81, NS).
Multifactor Comparisons. VA scores from the first
and second eyes tested were compared on the basis
of sex and the four racial categories in two separate
one-way ANOVAs to verify the possibility of differences
from any of the groups. Fourteen of the 16 possible
groups were tested; because there was only one child
in the Asian female group, this category could not
be included in the analysis. The ANOVA result was
significant for first eye tested (F6.623 = 2.65; P = 0.02),
second eye tested (F6623 = 2.47; P— 0.02), and binocular data (F6.644 = 2.49; P = 0.02). However, the post
hoc Scheffe test did not reveal the source of significance in any of these cases: no two groups were significantly different at the .05 level.
Binocular Versus Monocular Results. Mean differ-
ence between binocular and monocular VAs were
compared across the 19 age groups. Differences were
not significant for all groups (F lg623 = 1.04, NS).
Refractive Errors
The incidence and type of refractive errors found in
the 646 subjects was almost always identical for the
two eyes, as shown by the calculated correlations for
the whole sample, which were based on spherical
662
Investigative Ophthalmology & Visual Science, March 1995, Vol. 36, No. 3
0)
•o
64
32
16
8
4
2
1
0.5
0.25
0.13
12
18
24
30
36
30
36
Corrected age (months)
O)
64
32
16
8
4
2
1
0.5
0.25
0.13
12
18
24
Corrected age (months)
FIGURE l. (A) Mean binocular acuity (±1 SEM) and tolerance limits for 90% of the population with 95% probability based on data from 646 healthy infants and children as a function
of age. (B) Mean monocular acuity (±1 SEM) and tolerance limits for 90% of the population
widi 95% of probability based on data from 624 healthy infants and children as a function
of age.
equivalents (r = 0.97; P < 0.05). Refractive errors
ranged from —0.50 to +4.75 D in spherical equivalent
(mean = +1.25; SD = 1.18) D). Most eyes were hyperopic (49.5%) and had compound hyperopic astigmatism (23.8%); 8% had emmetropia. Other types of
refractive errors including myopia, hyperopic astigmatism, myopic astigmatism, compound myopic astigmatism, and mixed astigmatism. Each occurred in fewer
than 5% of the subjects.
No correlation was found between refractive error
and visual acuity. The Pearson coefficients for this
comparison in OD and OS were, respectively, r =
-0.18 for OD and r = -0.15 for OS (NS).
Testability and Test Time
Testability and test time were based on the 726 children from the initial sample. Binocular testability was
99.3%, and monocular testability was 96.1%.
Average test time of the entire sample, with the
exception of the youngest age group, was 12.61 minutes (SD = 5.28 minutes), with a median of 11 minutes and a mode of 10 minutes. This test duration
included one binocular and two monocular measurements, in addition to eye patching time. Minimum
mean test time (minutes) was 11.42 ± 3.55 for the 26month-old group, and maximum was 15.88 ± 7.83 for
the 2-month-old group.
The 2-week-old group was tested only binocularly,
and its average test time was 6.91 ± 3.63 minutes.
Agreement Among Testers
Monocular VA results in the TAC obtained for each
of the eight individual testers in the 646 children of
the final sample are shown in Figure 3 along with the
preliminary norms.44 To evaluate agreement among
observers, a two-way ANOVA including age as a factor
was performed combining data across the 20 age
groups, pooled into 7 age groups, according to similar-
•
>
Age Norms for Vistech-Teller Acuity Cards
663
64
32
16
8
4
2
1
0.5
0.25
0.12
TOLERANCE UMTS
- • - - PRELIMINARY NORMS
12
18
24
30
36
Corrected age (months)
B
64
32
16
8
4
2
1
0.5
0.25
0.1:
_ TOLERANCE UMTS
- - PRELIMINARY NORMS
12
18
24
30
36
Corrected age (months)
FIGURE 2. (A) Binocular tolerance limits of grating VA for 90% of the population with 95%
probability, measured by Vistech-Teller Acuity Cards, as a function of age. The preliminary
norms for the same instrument are presented for comparison (dashed lines). (B) Monocular
tolerance limits of grating VA for 90% of the population with 95% probability, measured
by Vistech-Teller Acuity Cards, as a function of age. The preliminary norms for the same
instrument are presented for comparison (dashed lines). VA = visual acuity.
ity in mean acuity. For binocular data, age groups
were: 0.5, 2, 3, 4, 6 to 18,20 to 24, and 26 to 36 months
old. The monocular age groups were: 2, 3, 4, 6, 8, 10
to 24, and 26 to 36 months old. The results for both
binocular (F28.645 = 1.17 NS) and monocular (F29.623
= 1.37, NS) VAs showed no significant differences
among testers for any of the age groups.
Test-Retest
Retests were performed only in special cases (see
Methods). Twenty-six boys and 23 girls, with corrected
age from 1 to 36 months, were uncooperative and had
low scores or high IADs. They represent 7.6% of the
total sample (49/646).
Monocular VA pairs of test and retest scores were
obtained in 49 children (98 eyes). Visual acuity in the
retest was higher in 38 pairs (39%), lower in 36 (37%),
and identical to that of the test in the remaining 24
(24%). Differences in VA between test-retest ranged
from 0 (N = 24) to 3 octaves (N = 1).
The percentage of acuity scores differing by a
given number of octaves in the 98 monocular testretest pairs is presented in Figure 4. Differences >1
octave were more frequently observed in children
from 0 to 6 and 19 to 24 months of age. In these, the
differences were approximately 35% of the retested
children.
Interocular acuity differences (IAD), or OD minus
OS, in the test were compared with those in the retest.
The IAD in the test ranged from 0 (N = 2) to 3 octaves
664
Investigative Ophthalmology & Visual Science, March 1995, Vol. 36, No. 3
TABLE 3. Inteocular Acuity Differences and Respective Tolerance Limits for the Vistech
Teller Acuity Cards
Absolute
Age
(months)
2
3
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
1
N
8
21
25
30
34
35
40
34
41
35
37
35
43
37
42
34
24
37
32
Mean
Difference
OD-OS
(act)
0.54
0.32
0.21
0.31
0.22
0.49
0.38
0.43
0.27
0.33
0.45
0.46
0.34
0.35
0.26
0.32
0.43
0.26
0.20
Tolerance Limits (oct) *
Absolute
SD (oct)
Mean
Difference
OD-OS
(oct)
SD
Difference
(oct)
OD Better
OS Better
0.78
0.46
0.31
0.29
0.31
0.48
0.44
0.55
0.30
0.38
0.50
0.45
0.47
0.46
0.29
0.37
0.39
0.3
0.32
0.06
0.27
-0.01
0.04
0.01
0.16
-0.05
0.12
-0.01
-0.04
0.02
0.01
-0.15
0.01
0.05
-0.05
0.08
-0.1
-0.03
0.97
0.49
0.38
0.43
0.39
0.67
0.58
0.69
0.41
0.5
0.68
0.65
0.56
0.59
0.38
0.49
0.58
0.38
0.38
2.72
1.32
0.83
0.96
0.83
1.56
1.14
1.57
0.83
1.01
1.43
1.36
0.99
1.23
0.83
0.98
1.37
0.69
0.78
2.6
0.78
0.85
0.88
0.81
L.24
L.24
.33
().85
L.09
L.39
.34
1.29
1.21
0.72
1.08
1.21
0.89
0.84
In the first two age groups, tolerance limits are calculated for 90% of the population with 90% probability.
(N = 1), and in the retest they ranged from 0 (N =
21) to 1.5 octaves (N = 5). In 34/49 (70%) children,
IADs were lower in the retest.
Confidence of Grating Acuity Estimates
The subjective degree of confidence in each grating
acuity judgment was scored by each tester in a 5-point
scale in 95% of the tests. Excellent (86%) and good
(12%) confidence was reached in the majority of binocular and monocular measurements. The remaining
2% were classified as fair.
norm statistics and its adequacy for clinical interpretation of VA and IAD findings. We compare our findings
with those of previous normative studies, mostly with
preliminary norms.44 The norms derived by Mayer et
al (this issue, page 671) from a sample with a very
different socioeconomic composition are compared
to the present norms. Complementary aspects concerning data (binocular versus monocular VA, refractive errors, and VA) or procedure variables
(agreement among testers, test-retest in special cases,
binocular versus monocular VA, testability, and test
duration) are discussed.
DISCUSSION
Tolerance Limits as Norm Statistics
The present study provides binocular and monocular
tolerance limits for grating acuity measured with the
Vistech-Teller Acuity Cards in a large sample of
healthy children from 0 to 36 months of age. Norms
for interocular acuity differences are also presented.
The study showed that testers can be trained to measure VA within a reasonably close agreement and that
a more stringent procedure than that typically used
with the TAC44 can be used with short testing times
compatible with clinical or screening purposes. This
is the first study to provide both binocular and monocular norms in a large sample of infants and children
primarily from low-income families of different racial
origins.
Below we discuss the use of tolerance limits as
Visual acuity estimates can vary for the same age in
different children and can also vary in the same child
in different measurements. The first variability is inherent in the function being measured, and the second is inherent in the precision with which it can be
measured. Assuming both are normally distributed, it
can be said that a given result is normal if it falls within
the interval delimited by standard deviations above
and below the mean. This interval represents the confidence limits with which the mean obtained can be
described. Confidence limits cover a population parameter (the mean) with a certain confidence. This
can be translated to the present situation in this manner: If the VA value is within the confidence interval,
there is a defined probability (90%, 95%, 98%, or
Age Norms for Vistech-Teller Acuity Cards
6
12
665
II
24
normal" cases in the interval. In the limit, if a 99.9%
tolerance interval was chosen, almost any VA would
have been classified as "normal." The objective of
defining norms is to draw a line that will exclude the
pathologies that exist in the population.
In the present study, the purpose was to establish,
with a certain probability, intervals of normal values
of VA for a specified percentage of the population.
The interval chosen here, which covers 90% of the
population, excludes 5% above the upper limit and
5% below the lower limit. In practice, for clinical purposes, it is the 5% below the lower limit that are of
concern to the clinician. It is evident that the choice
of a tolerance interval will depend on the number of
referrals one can afford or find profitable. It affects
the balance between false-positive and false-negative
results. Tolerance limits were also used by Heersema
and van Hof-van Duin29 in their normative study of
VA in toddlers. In that study, they showed that VA is
an approximately linear function of the logarithm of
age. However, there might be clinical situations in
which this condition is not met by the data. Because
tolerance limits do not depend on an assumption of
linearity with relation to age, they would be applicable
in these cases.
M
Comctad «gt (months)
64
MM
32
16
4
2 8*_^-- *
1
0.S
0J5
1/
l
0.13
6
12
18
24
30
I
OD r*•
6
12
II
24
11
II
24
30
Cometh *g« |modth«)
Corrected «8« (months)
^
30
6
Comcttd age (months)
12
It
24
30
Corrected age (months)
64
92
16
5P *4
I 2
p./-
O.S
0.2S
RO
^ tr
— — "
1
"
*•"
1 /1
0.13
6
12
1>
24
30
Comcttd ige (monthi)
36
6
12
11
24
30
36
Comcttd age (months)
FIGURE 3. Individual monocular grating VA scores obtained
with the Vistech-Teller Acuity Cards by the eight testers,
identified by the initials in the upper right corner of each
graph. Dashed lines are preliminary VA monocular norms.
• = OD; O = OS. VA = visual acuity.
whatever one chooses to define) that it is accepted as
being equal to the mean.
One can, on the other hand, choose to obtain a
range of VA values that encompasses a fixed portion
(e.g., 70%, 90%) of the population with a specific
probability. These intervals are called tolerance intervals, and their extreme points are called tolerance
limits.46 In this case, the interval represents the normal
VA range for a defined percentage of the population
(any percentage that one defines). Translating the
concept to the present study, this means that if a given
VA value is within the tolerance limits fixed for 90%
of the population, it can be said, with a certain defined
probability, that the statement "VA is within normal
range" is true. The greater the portion of the population covered, the greater will be the number of "ab-
Present Norms and Other Normative Studies
The first norms available for TAC presented by Teller
et al25 are based on McDonald et al studies19"21 in the
case of monocular norms and are distributed with the
TAC.44
The finding of higher VAs than those presented
in the norms above has been the trend in recent stud3 29 48 49
ies, ' ' ' and Mayer et al, this issue, which show consistent results despite the use of different procedures.
It has been suggested29 that McDonald et al19"21 could
have underestimated the mean acuity because even
though photorefractional screening was performed, a
full ophthalmologic assessment was not.
A discussion of the causes for the higher VAs
found in the first year of life should take sample size
and composition into account. Even though sample
size was large, thus reducing the variance, and the
sample was likely to be homogeneous, as shown by the
absence of difference between any two groups when
tested with the Scheffe procedure, these factors could
not have elevated the averages. A small sample would
result in an irregular function, but not in a uniformly
lower (or higher) one. A larger sample simply contributes to a smoother function, widi lower variance
around each point.
Possible explanations for the higher VA scores
found in the 4- to 12-month group must be related to
procedural factors. The majority of the data here
(69%) were collected in an environment familiar to
the child—the day care unit. This might have contrib-
666
Investigative Ophthalmology & Visual Science, March 1995, Vol. 36, No. 3
100 r
0-6
7-12
13-18
19-24
25-36
Age groups (months)
FIGURE 4. Percentage of grating visual acuity scores differing by a given number of octaves
in the 98 monocular test-retest pairs. Number of pairs tested per age group is indicated
above the corresponding bar.
uted to the childrens' cooperation in the test situation.
Similarly, small changes in stimuli between the prototype cards used for the preliminary norms and the
Vistech-Teller cards might also account for the differences. The latter cards have larger fields, which might
be especially helpful to younger infants. Dobson et
al,50 however, reported no differences in VA for the
prototype and Vistech cards. It is, therefore, doubtful
that this explains the differences.
Norms for Interocular Acuity Differences
Knowledge of IAD is critical for the clinical diagnosis
of amblyopia and of monocular disorders. Birch and
Hale48 find that interocular difference offers a better
criterion for monocular acuity deficit, with higher sensitivity than the mere application of the norms. Differences found in the sample used for the present norms
are within the range of those found by others48 (Mayer
et al, this issue, page 671).
Comparison of Studies: Children From Middleto High- Versus Low-Income Families
As reported, this study was conducted in parallel
with a similar study conducted by Mayer et al (this
issue, page 000) .51~55 Each had as its objective the
establishment of norms for the TAC, and each used
a large number of subjects. Some differences between the two studies are number of testers, racial
and socioeconomic composition of the sample, testing procedures, number of age groups, and age intervals. The present study used eight testers in a low-
income sample, whereas Mayer et al used two testers
in a middle- to high-income sample. Racial differences between the two samples were also pronounced. Threshold determination differed in that
the Mayer et al study used the classical procedure, 44
whereas in this a modified staircase procedure was
adopted (see Methods).
Upper and lower limits of both sets of results
are compared in Figure 5. It is important to point
out how similar the lower limits are, despite the very
different composition of the samples. It should be
stressed that although children of low-income families composed most of the sample, they attended
public day care centers in Sao Paulo and were not
undernourished. Low socioeconomic level might be
responsible for lower stimulation, greater difficulty
in obtaining medical care when needed, and diminished primary medical care. These factors, if present, did not seem to have any effect on the visual
acuity of these children.
Complementary Aspects
Concerning Data. Binocular versus monocular VA.
Birch and Hale's48 report on forced-choice preferential looking shows slightly higher binocular VA scores
after 6 months, but the corresponding statistical analysis shows that the differences are not significant for
children from 13 to 60 months of age. The present
study finds no statistical difference between binocular
and monocular VA, but it also shows slightly higher
binocular VA after 3 months.
667
Age Norms for Vistech-Teller Acuity Cards
64
32
16
8
4
2
S;
1
ENT STUDY, PREDICTION LIMITS 2.SK
0.5
PRESENT STUDY, TOLERANCE LIMITS 6%
0.25
MAYER ET AL., PREDICTION LIMITS 2.5%
0.13
12
18
24
30
36
Corrected age (months)
FIGURE 5. Monocular grating VA prediction limits (95%) and tolerance limits (90% of the
population with 95% probability) as a function of age, calculated with data from the original
sample. Monocular grating VA prediction limits (95%) from the study by Mayer et al (this
issue, page 671) are also presented for comparison (dashed lines).
Refractive errors and VA. Not many studies of VA
determine refractive error. Ophthalmologic and orthoptic examination are an essential prerequisite of
normative studies because it is absolutely necessary
that any known pathology be excluded from the normal, sample. In the present study, these factors were
responsible for the exclusion from the original sample
of 14 children with refractive error below the cutoff
criteria (see Methods). In the sample used for the
norms (n = 646), refractive error, within the narrow
range accepted for the sample, was not correlated with
VA. This confirms other reports33'34 showing that refractive error is not causally related to the development of VA. Nevertheless a relationship might exist if
extreme cases had been included.
Concerning Procedural Variables. Testability and test
duration. Testability and test durations found in the
present study showed that VA assessment of most children could be completed in a short time. Binocular
and monocular testabilities reported in the literature
ranges from 89% to 100%.19"21 The testabilities found
here (99.3% for binocular and 96.2% for monocular
measurements) are at the high end of this range.
In the present study, test durations for three tests,
two monocular and one binocular, were 13 minutes
on average (SD = 5.66 minutes), and it could be estimated that each test took about 4.3 minutes. This
average refers to the whole sample, but it must be
kept in mind that younger subjects took a longer time
to be tested (—7 minutes for infants younger than 1
month and 5 minutes for 2- and 3-month-old infants.
When comparing test durations reported from different studies, the procedure used should be taken
into account. Test time is dependent on the psychophysical method. The modified staircase procedure
adopted here might introduce a more precise evaluation of the grating acuity threshold at the cost of
lengthened test time. In fact, test times found by VitalDurand and Hullo56 using TAC were about half as
long as ours. Times found by other authors using
TACzll25'3° were slightly shorter but more comparable
to ours, despite the fact that they did not use a staircase
procedure. The high success rate also contributed to
the lengthening of the average test duration because
every child was tested regardless of how long it took.
Agreement among testers. Screening programs re-
quire large groups of testers to be effective within
budget and time constraints. One of the concerns of
this work was to find out whether we could train testers
to produce data with good intertester agreement. The
largest numbers of testers reported in single studies
of acuity card measurements were, respectively, fiveD?
and four,08 in two studies using infants with perinatal
complications; agreement was similar to that reported
previously but was not as good as found for normals.
Quinn et al,28 who used three testers in a clinical setting, report significant intertester differences and recommend, in line with the two former studies, that
interobserver reliability be tested when using TAC.
The eight testers used in the present study were
carefully trained before the beginning of data collection. This, and the continuous monitoring of the tests
by one of us (SS), was responsible for the fact that no
significant differences in average acuity values were
found among testers. It must be pointed out that the
comparison performed here is not very sensitive and
668
Investigative Ophthalmology & Visual Science, March 1995, Vol. 36, No. 3
that a better approach would have been to collect
data from the same eyes. However, it would have been
difficult to submit the same child to eight different
testers.
Test-retest in special cases. Retests, performed in
special cases, showed higher VAs about as often as
there were of lower VAs (39% versus 37%) and equal
scores in the remaining cases. The reduction of interocular differences found in 70% of the cases indicates that the retest scores might be more reliable.
However, because retests were only performed in special cases, the tester might have taken extra care knowing that something went wrong on the first test. In
addition, reduction of interocular differences might
be due to the phenomenon of regression of the
mean—the tendency for normal subjects with extreme values to have less extreme values on retest. The
implication is that VAs in the test might have been
underestimated as frequently as overestimated, which
is to be expected because these children were uncooperative, and testers reported low confidence in their
estimates.
Because test-retests were performed only in especially difficult cases, it is not appropriate to analyze
the results as representing agreement among testers.
A measure of agreement depends on the assumption
that the child does not change from test to retest.
The expectation of the testers was the opposite—to
achieve better cooperation or confidence on retest.
CONCLUSIONS
The present norms provide binocular and monocular
data mostly from infants and children of low-income
families, whereas Mayer et al's study (this issue, page
671) provides monocular norms for children primarily
from middle- to high-income families. Regardless of
these differences and of the fact that in the present
study a more stringent procedure for VA estimates was
used, the norms obtained in both studies are remarkably similar. The two studies, therefore, complement
each other and strongly point to the need for adoption
of new norms for the Vistech-Teller Acuity Cards.
Norms for interocular acuity differences, which are of
important clinical application, are also reported.
In addition, we suggest that the earliest phase of
development be focused in future work. In this phase,
the growth in VA is rapid, and it would be convenient
to have binocular and monocular norms for Vistech Teller Acuity Cards based on shorter intervals. Testing
of a larger number of children during the first 3
months of life would be required for the completion
of an adequate set of norms. It should also be noted
that binocular data may be clinically valuable, mainly
in the first month, during which it is difficult to obtain
monocular measurements.
Key Words
infant vision, grating visual acuity, visual acuity development,
Vistech-Teller Acuity Cards, visual acuity age norms
Acknowledgments
The authors thank Velma Dobson, PhD, for her constant
support throughout the study and for her visit to Brazil at
the beginning of this study. The authors also thank Davida
Teller, PhD, for her critical review of and suggestions for
this manuscript; Jair Licio Ferreira Santos, PhD, from the
Public Health Faculty of the University of Sao Paulo for his
competent advice on all aspects of the statistical treatment
of the data; Adriana Berezovsky, Eny Gitelman, Marcia Maia,
M. Valeria Ferrari, Raquel Eliezer for their collaboration in
performing orthoptic evaluations and testing the children;
Fernando Cunha and Regis Oliveira for testing the children;
and of Emilio de Haro-Munoz, MD, and Cesar Lipener, MD,
for performing ophthalmologic examinations.
References
1. Dobson V, Teller DY. Visual acuity in human infants:
A review and comparison of behavioral and electrophysiological studies. Vision Res. 1978; 18:1469-1483.
2. Boothe RG, Dobson V, Teller DY. Postnatal development of vision in human and nonhuman primates.
Ann Rev Neurosci. 1985;8:495-545.
3. van Hof-van Duin J. The development and study of
visual acuity. Dev Med Child Neurol. 1989; 31:543-552.
4. Dobson V. Behavioral assessment of visual acuity in
infants. In: Berkley MA, Stebbins WC, eds. Comparative
Perception. Vol. 1. Basic Mechanisms. John Wiley;
1990:487-521.
5. Jacobson SG, Mohindra I, Held R. Development of
visual acuity in infants with congenital cataracts. BrJ
Ophthalmol. 1981;65:727-735.
6. Birch EE, Stager DR, Wright WW. Grating acuity development after early surgery for congenital unilateral
cataracts. Arch Ophthalmol. 1986, 104:1783-1787.
7. Birch EE, Stager DR. Prevalence of good visual acuity
following surgery for congenital unilateral cataracts.
Arch Ophthalmol. 1988; 106:40-43.
8. Mayer DL, Moore B, Robb RM. Assessment of vision
and amblyopia by preferential looking tests after early
surgery for unilateral congenital cataracts. / Pediatr
Ophthalmol Strabismus 1989;26:61-68.
9. Birch EE, Stager DR, Berry P, Everett ME. Prospective
assessment of acuity and stereopsis in ambylopic infantile esotropes following early surgery. Invest Ophthalmol
Vis Sri. 1990; 31:758-765.
10. Wiesel TN. Postnatal development of visual cortex and
the influence of the environment. Nature. 1982;
299:583-591.
11. Ikeda H, Wright MJ. A possible neurophysiological
basis for amblyopia. Br OrthopJ. 1975;32:2-13.
12. Teller DY, Morse R, Borton R, Regal D. Visual acuity
for vertical and diagonal gratings in human infants.
Vision Res. 1974;14:1433-1439.
13. Gwiazda J, Brill S, Mohindra I, Held R. Infant visual
acuity and its meridional variation. Vision Res. 1978;
18:1557-1564.
Age Norms for Vistech-Teller Acuity Cards
14. Atkinson J, Braddick O, Pimm-Smith E. 'Preferential
looking' for monocular and binocular acuity testing
of infants. BrJ Ophthalmol. 1982;66:264-268.
15. Mayer DL, Dobson V. Visual acuity development in
infants and young children as assessed by operant
preferential looking. Vision Res. 1982;22:1141 —1151.
16. Birch EE, GwiazdaJ, Bauer JA Jr., Naegele J, Held R.
Visual acuity and its meridional variations in children
aged 7-60 months. Vision Res. 1983;23:1019-1024.
17. Lewis TL, Maurer D. Preferential looking as a measure
of visual resolution in infants and toddlers: A comparison of psychophysical methods. Child Dev. 1986;
57:1062-1075.
18. Fantz RL. Pattern'Vision in young infants. Psychol Rec.
1958;8:43-47.
19. McDonald MA, Dobson V, Sebris SL, Baitch L, Varner
D, Teller DY. The acuity card procedure: A rapid test
of infant acuity. Invest Ophthalmol Vis Sri. 1985;
26:1158-1162.
20. McDonald MA, Mohn G, Teller DY, Dobson V. Monoclear acuity in normal infants: The acuity card procedure. Am] Optom Physiol Opt. 1986;63:127-134.
21. McDonald MA, Ankrum C, Preston K, Sebris SL. Monocular and binocular acuity estimation in 18- to 36month-olds: Acuity card results. Am] Optom Physiol Opt.
1986; 63:181-186.
22. Dobson V, Kohl P, Stern N, Samek M, Preston K.
Visual acuity screening of infants and young children
with the acuity card procedure. / Am Optom Assn.
1986;27:284-289.
23. Yamamoto M, Brown AM. Vision testing of premature
infants using grating acuity cards. Fol Ophthalmol Jap.
1985; 36:796-799.
24. Mohn G, van Hof-van Duin J. Rapid assessment of
visual acuity in infants and children in a clinical setting, using acuity card. Doc Ophthalmol Proc Ser.
1986; 45:363-372.
25. Teller DY, McDonald MA, Preston K, Sebris SL, Dobson V. Assessment of visual acuity in infants and children: The acuity card procedure. Devel Med Child Neurol. 1986;28:779-789.
26. Mahon K. Preferential looking and acuity card testing
in the private practitioner's office. Am Orthop J.
1987; 37:86-94.
27. Trueb L, EvansJ, Hammel A, Bartholomew P, Dobson
V. Assessing visual acuity of visually impaired children
using the Teller acuity card procedure. Am Orthop J.
1992; 42:149-154.
28. Quinn GE, Berlin JA, James M. The Teller acuity card
procedure. Three testers in a clinical setting. Ophthalmology. 1993; 100:488-494.
29. Heersema DJ, Van Hof-Van Duin J. Age norms for
visual acuity in toddlers using the acuity card procedure. Clin Vision Set. 1990;5:167-l74.
30. Courage ML, Adams RJ. Visual acuity assessment from
birth to three years using the acuity card procedure:
Cross-sectional and longitudinal samples. Optom Vis
Sti. 1990; 67:713-718.
31. Chandna A. Natural history of the development of
visual acuity in infants. Eye 1991;5:20-26.
32. Preston KL, Freedman HL. The Teller acuity card
669
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
procedure: An acuity screening technique for pre-verbal children. Am Orthop J. 1988; 38:24-32.
Kohl P, Rolen RD, Bedford AK, Samek M, Stern N.
Refractive error and preferential looking visual acuity
in human infants: A pilot study. J Am Optom Assn.
1986;57:290-296.
Kohl P, Samek MS. Refractive error and preferential
looking visual acuity in infants 12-24 months of age:
Year 2 of a longitudinal study. J Am Optom Assn.
1988;59:686-690.
Moseley MJ, Fielder AR, Thompson JR, Minshull C,
Price D. Grating and recognition acuities of young
amblyopes. Br J Ophthalmol. 1988; 72:50-54.
Katz B, Sireteanu R. Der Teller-Acuity-Card-Test:
Moglchkeiten und Grezen der klinischen Anwendung. Klin Mbl Augenheilk. 1989; 195:17-22. In German.
Sireteanu R, Fronius M, Katz B. A perspective on psychophysical testing in children. Eye. 1990;4:794-801.
Dobson V, Quinn GE, Biglan AW, Tung B, Flynn JT,
Palmer EA. Acuity card assessment of visual function
in the cryotherapy for retinopathy of prematurity trial.
Invest Ophthalmol Vis Sti. 1990;31:1702-1708.
Dickey CF, Metz HS, Stewart SA, Scott WE. The diagnosis of amblyopia in cross-fixation. J Pediatr Ophthalmol Strabismus. 1991;28:171-175.
Boergen KP, Kau T, Lorenz B, Rosenauer E. Untersuchungen zur normalen und gestorten visuellen Entwicklung mit Preferential Looking (PL). Klin Mbl
Augenheilk. 1991; 199:103-109. In German.
Hertz BG. Acuity card testing of retarded children.
Behav Brain Res. 1987;24:85-92.
Hertz BG. Use of acuity card method to test retarded
children in special schools. Child Care Health Dev.
1988; 14:189-198.
IBGE. Censo demogrdfico: dados gerais, migracdo, instrufdo, fecundidade, mortalidade. Rio de Janeiro, RJ: Fundacao Instituto Brasileiro de Geografia e Estatistica IBGE; 1983;l(4):34-35. In Portuguese.
Teller Acuity Card Handbook. Vistech Consultants Inc,
Dayton Ohio, 1989.
Dobson V, Salem D, Carson JB. Visual acuity in infants—The effect of variations in stimulus luminance
within the photopic range. Invest Ophthalmol Vis Sci.
1983;24:519-522.
Dixon WJ, Massey FJr. Introduction to statistical analysis. McGraw-Hill; 1957:130.
Rosner B. Statistical methods in ophthalmology: An
adjustment for the intraclass correlation between eyes.
Biometrics. 1982; 38:105-114.
Birch EE, Hale L. Criteria for monocular acuity deficit
in infancy and early childhood. Invest Ophthalmol Vis
Sri. 1988; 29:636-643.
Mayer DL, Warner A, Beiser A, Lang J, McNiff K.
Normative acuity card study: I: Monocular norms.
ARVO Abstracts. Invest Ophthalmol Vis Sri. 1992;
33:1353.
Dobson V, Luna B. Prototype and Teller acuity cards
yield similar acuities in infants and young children
despite stimulus differences. Clin Vision Sri. 1993;
8:395-400.
670
Investigative Ophthalmology & Visual Science, March 1995, Vol. 36, No. 3
51. Raye K, Pratt E, Beiser A, Lang J, Mayer DL. Normative
Teller acuity card study: II: Test-retest reliability.
ARVO Abstracts. Invest Ophthalmol Vis Sd. 1992; 33:716.
52. Warner AF, Beiser A, Lang JM, Mayer DL. The normative acuity card study: Study design and recruitment.
ARVO Abstracts. Invest Ophthalmol Vis Sd. 1991;32:962.
53. Salomao SR. Acuidade de resolufdo e mudanfas no erro
refrativo em criangas ale 3 anos de idade. Sao Paulo, Brasil:
Escola Paulista de Medicina; 1990. Master's thesis. In
Portuguese.
54. Salomao SR. Acuidade visual de resolucao medida
pelo teste dos cartoes de acuidade: Normas populacionais e aplicacoes em triagem e na clunica. Sao Paulo,
Brasil: Escola Paulista de Medicina; 1993. Doctoral
thesis. In Portuguese.
55. Salomao SR, Ventura DF. The Brazilian large scale
acuity card study: I: The normative study. ARVO Abstracts. Invest Ophthalmol Vis Sd. 1993;34:1420.
56. Vital-Durand F, Hullo A. La mesure de l'acuite visuelle du nourisson en six minutes: Les cartes d'acuite de Teller. J Fr Ophthalmol. 1989;3:221-225. In
French.
57. Mash C, Dobson V, Carpenter N. Interobserver
agreement for measurement of grating acuity and interocular acuity differences with the Teller Acuity
Card procedure. Vision Res. 1994;35:303-312.
58. Dobson V, Carpenter NA, Bonvalot K, Bossier J. Interobserver agreement in infants with perinatal complications. Clin Vision Sd. 1990;6:39-48.
T