Stereoacuity norms for school-age children usingthe Frisby stereotestPamela M. Anketell, MMedSci, BSc Orth, Kathryn J. Saunders, PhD, BSc MCOptom, andJulie-Anne Little, PhD, BSc, MCOptom

BACKGROUND The Frisby stereotest and the TNO test for stereoscopic vision are popular clinical tests

Author affiliations: Vision Sciencof Ulster, Cromore Road, ColeraFunded by the Northern Irela

Office Northern Ireland (EAT/4Submitted March 4, 2013.Revision accepted August 20,Correspondence: Pamela M. A

Group, School of Biomedical ScieCo Londonderry Northern IrelanCopyright � 2013 by the Am

Strabismus.1091-8531/$36.00http://dx.doi.org/10.1016/j.ja

582

for assessing stereoacuity: however, reference data for school-age children for the Frisbystereotest are limited. This study compared stereoacuity results of both tests in a largesample of typically developing school-age children.

METHODS Primary (elementary school grades 1-6) and post-primary (high school grades 7-11) stu-

dents aged 6-16 years were recruited and assessed in schools. Stereoacuity thresholdswere measured using the Frisby and TNO stereotests. Children with ocular pathology,anisometropia of $1.00 D, interocular difference of visual acuity $0.2 logMAR, or stra-bismus were excluded.

RESULTS A total of 212 children were recruited; data for 186 subjects were analyzed. Median Frisby

stereoacuity scores were, for crossed disparity, 20 arcsec for primary and 10 arcsec forpost-primary children and, for uncrossed disparity, 25 arcsec (primary) and 10 arcsec(post-primary). TNO stereoacuity was 60 arcsec for both age groups. For Frisby stereoa-cuity, scores of 85 arcsec (crossed) and 170 arcsec (uncrossed) or better were achieved by95% of primary school children; scores of 85 arcsec (crossed and uncrossed) or betterwere achieved by 95% of post-primary subjects. A statistically significant difference inmedian stereoacuity scores was noted across age groups for the Frisby stereotest (crossed:z 5 4.67, P\0.0001; uncrossed: z 5 4.67, P\0.0001). No statistically significant differ-ence in stereoacuity scores was noted with the TNO stereotest (z5 1.35, P5 0.18). A sig-nificant weak correlation was found between the Frisby and TNO stereotests (Frisby[crossed], r 5 0.21 P\ 0.005).

CONCLUSIONS These data describe normative values for the Frisby stereotest for children aged 6-16 years.

Participants recorded significantly better stereoacuity scores with the Frisby stereotestthan the TNO stereotest. The Frisby stereotest values are weakly correlated with theTNO stereoacuity test. ( J AAPOS 2013;17:582-587)

Stereoacuity assessment may record a “pass” or “fail”or it can be quantified for use as an indicator oftreatment outcomes, for example, in amblyopia

management.1-3 Stereoacuity can be assessed as crossed(target appears to be in front of the reference plane) oruncrossed (target appears to be behind or sink into thereference plane) disparity using the Frisby stereotest(Haag Streit UK Ltd [Clement Clarke InternationalLtd], Harlow, Essex). Shipmann and Cohen4 reportedthat crossed stereoacuity is better compared to uncrossed

e Research Group, School of Biomedical Sciences, Universityine, Northern Irelandnd Health and Social Care Research and Development197/09).

2013.nketell, MMedSci, Bsc Orth, Vision Science Researchnces, University of Ulster, Cromore Road, Coleraine,d BT52 1SA (email: anketell-p@email.ulster.ac.uk).erican Association for Pediatric Ophthalmology and

apos.2013.08.012

in exophoric patients, whereas uncrossed scores are betterthan crossed in esophoric patients.

The TNO test of stereoscopic vision (Lam�eris OotechBV, Nieuwegein, Netherlands) is also commonly used toassess stereoacuity. The TNO is an anaglyph stereotestpermitting assessment of stereoacuity without monocularcues5,6; it does, however, require the use of red-green spec-tacles that may limit cooperation in young children orthose with sensory issues. The Frisby stereotest is a real-depth free space stereotest that does not require spectaclesand can measure a wide range of disparities. Success ofmeasuring stereoacuity using the Frisby stereotest in chil-dren is well documented.7 However, if the head movesduring testing, a patient could perceive monocular cuesthat could falsely elevate the stereoacuity score.6 Mannyand colleagues.8 assessed stereoacuity using the Frisby ster-eotest in children with normal and abnormal binocularvision to determine whether monocular cues could beused to identify disparity. They found that no child withstrabismus was able to achieve a stereoacuity score.8 How-ever, Cooper and Feldman9 reported that undermonocular

Journal of AAPOS

Table 1. Frisby stereotest scores in arcsec (seconds of arc) at eachtest distance

Test distance, cm 6 mm 3 mm 1.5 mm

30 600 300 15040 340 170 8550 215 110 5560 150 75 4070 110 55 3080 85 40 20150 25 10 5

Volume 17 Number 6 / December 2013 Anketell, Saunders, and Little 583

viewing conditions a stereoacuity score was recorded inadult participants.Reference data describing the non-Gaussian distribution

of stereoacuity scores for Randot stereotests, for example,has been well described.10 However, reference data forthe Frisby stereotest for school-age children is limited.Heron and colleagues11 reportedmedian stereoacuity scoresranging between 27.5 and 16.5 arcsec for children aged 3-7years (n 5 30-37 per age group). Leat and colleagues12

reported mode scores of 20 arcsec for 32 children acrosstwoagegroups (5-8years and8-20years). Simons13 reportedthat 75% of 38 children achieved a Frisby stereoacuity scoreof 250 arcsec. Normative data for the TNO stereotest areavailable for large populations, including preschool13,14

and school-age 5,3,15 children as well as adults.11

The relationship of stereoacuity scores between randotand real-depth stereotests has been well described12,16-18;however, review of the literature reveals wide variationin the correlation of stereoacuity scores comparing theTNO and Frisby stereotests.13,6,11The purpose ofthe present study was to provide normative crossed anduncrossed data for the Frisby stereotest for school-agechildren and to explore how the Frisby stereoacuityscores relate to TNO stereoacuity scores.

Subjects and Methods

This study received ethical approval from the University of Ulster

Research Ethics Committee and adhered to the tenets of the

Declaration of Helsinki. As part of a wider study of vision, partic-

ipants were recruited from primary (elementary grades 1-6) and

post primary (high school grades 7-11) schools from one region

of the United Kingdom across four age groups; 6-7 years, 9-10

years, 12-13 years, and 15-16 years. The schools represented a

cross section of academic ability and socioeconomic factors. All

participants were tested in school during the school day. Written

informed consent was obtained from the parents or guardians of

children. At the time of consent, parents or guardians were asked

to complete a short list of questions relating to their child’s

previous ocular history, family history, and general health. Data

from subjects with an interocular visual acuity difference of

$0.2 LogMAR and/or presence of strabismus and/or anisome-

tropia of $1.00 D were excluded.

The Frisby stereotest consists of three perspex plates of varying

thickness (6 mm, 3 mm, 1.5 mm). Real depth is achieved by print-

ing elements of the target on both sides of the plate. The test

allows measurement of crossed and uncrossed disparities by

reversing the plate presentation. The 6 mm test plate was initially

presented to participants at a distance of 80 cm. If the participant

could reliably identify the “circle” in two of three presentations,

the 3 mm plate was presented and the process repeated. If the

3 mm test plate was correctly identified the assessment was

repeated for the 1.5 mm test plate. If all three test plates were

correctly identified at 80 cm, the test distance was increased to

150 cm and the process was repeated for each test plate, starting

with the 3 mm test plate, until threshold was reached. For those

unable to identify the circle on the 6 mm test plate at 80 cm the

Journal of AAPOS

disparity was increased by moving the plate closer to the partici-

pant in 10 cm increments until the circle was identified. If the

circle was not identified on the 6 mm test plate 30 cm distance,

a fail was recorded. If the circle was identified for at least two of

three presentations on the 6 mm test plate the 3 mm and

1.5 mm test plates were assessed until threshold was achieved at

this distance. The stereoacuity scores for each test distance are

described in Table 1.

Crossed and uncrossed threshold disparities were measured

following this method. The stereoacuity score recorded was the

best stereoacuity score for which at least two of three observations

were correctly identified. Monocular cues were minimized by

presenting the plate against a white background, keeping the

test plate stationary in the participant’s visual axis, and discour-

aging head movement with prior instructions and reminders

throughout the test.

The TNO stereotest uses an anaglyphic technique, presenting

stereoscopically hidden images only visible through dissociative

red-green spectacles. The test was presented at 40 cm according

to the manufacturer’s instructions.19 If the screening plates were

correctly identified, the quantification plates were presented to

participants. The participant was asked to identify the “missing

piece of pie” in each circle. At each stereoacuity level offered by

the TNO stereotest there are two test targets; stereoacuity was

recorded as the best stereoacuity score at which both targets

were correctly identified.

The order of stereotest assessment varied with neither test

always presented first by the examiner. The order of presentation

of the crossed and uncrossed targets using the Frisby was also

random. For all stereotests, participants were tested wearing their

habitual spectacle correction.

Monocular recognition acuity (distance, LogMAR Crowded

Acuity Test [Keeler, Windsor, UK]; near, Massachusetts

HOTV Near Visual Acuity Test [Good Lite, Elgin, IL]), cover

test, and ocular motility were recorded for each participant.

Cycloplegic autorefraction was performed 30 minutes after instil-

lation of 1% cyclopentolate hydrochloride (NVision-K5001

open-field autorefractor; Shin Nippon, Tokyo, Japan).20

Results

A total of 212 participants were recruited. Of these, 3 wereabsent from school during data collection and were nottested, and 4 were excluded on the basis of parent-reported medical or ocular history. The remaining205 participants successfully completed the Frisby and

Table 2. Median and interquartile ranges for Frisby stereotest and the TNO stereotest for each age group and the post-primary groups

Group Sample size

Frisby stereoacuity, arcsec

TNO stereoacuity, arcseccrossed uncrossed

Median (IQR) 95th centile Median (IQR) 95th centile Median (IQR) 95th centile

6-7 years 46 25 (20-40) 75 25 (20-40) 85 60 (60-120) 4809-10 years 51 20 (10-40) 85 20 (20-40) 215 60 (60-120) 48012-13 years 47 10 (5-25) 40 10 (5-25) 55 60 (60-120) 24015-16 years 42 10 (5-20) 85 10 (10-25) 85 60 (60-60) 480Primary age groupa 97 20 (20-40) 85 25 (20-40) 170 60 (60-120) 480Post primary age groupb 89 10 (5-20) 85 10 (10-25) 85 60 (60-120) 480

Arcsec, seconds of arc; IQR, interquartile range.a6-7 years and 9-10 years, inclusive.b12-13 years and 15-16 years, inclusive.

FIG 1. A, Box plot depicting Frisby (crossed) stereoacuity scores foreach age group. Median is indicated by a line, the edges of the boxrepresent the IQR, x represents the range and and the whiskerlimits the 5th and 95th centile. B, Box plot depicting Frisby (crossed)stereoacuity scores for each age group. Median is indicated by a line,the edges of the box represent the IQR, x represents the range andand the whisker limits the 5th and 95th centile.

584 Anketell, Saunders, and Little Volume 17 Number 6 / December 2013

TNO stereotests. Data were omitted for an additional19 children due to an interocular visual acuity differenceof $0.2logMAR and/or anisometropia $1.00 D and/orstrabismus; 186 study subjects (103 males [55%]) wereanalyzed. Of these 186 subjects, 7 (3.8%), ranging in agefrom 6 to 15 years of age, who were unable to correctlyidentify both responses at the level of 480 arcsec wereclassed as failing the TNO stereotest.

Stereoacuity scores were not normally distributed.Nonparametric statistics were used to analyze the rawscores and parametric analyses were used to analyze thelogarithm of the raw scores. Outcomes were the samewhether the parametric or nonparametric analyses wereused; therefore, nonparametric analyses of the raw stereo-acuity scores are presented. No statistically significant dif-ference was noted in stereoacuity scores between boys andgirls for either test (P. 0.4 for both).Median, interquartilerange, and 95th centile (upper limit of normative range)stereoacuity scores are presented for each age group inTable 2.

Kruskal-Wallis rank test (K) was used to analyze stereo-acuity scores across the four age groups. Frisby (crossedand uncrossed) stereoacuity scores were statistically sig-nificantly different across age groups (crossed: K 522.72, P \ 0.0005; uncrossed: K 5 22.83, P \ 0.0005).Post hoc analysis demonstrated no statistically significantdifference inmedian stereoacuity score between age groups6-7 years and 9-10 years (Mann-Whitney test; Frisby[crossed]: z 5 1.39, P 5 0.17; Frisby [uncrossed]: z 51.32, P5 0.19) and age groups 12-13 years and 15-16 years(Mann-Whitney; Frisby crossed; z 5 �0.64 , P 5 0.52,Frisby uncrossed; z 5 �0.81, P 5 0.42). No statisticallysignificant difference in median stereoacuity scores wasfound across age groups for the TNO test (K 5 1.49,P 5 0.68). The distribution of Frisby (crossed anduncrossed) stereoacuity scores is shown in Figure 1A-B.

As significantly better Frisby stereoacuity scores werenoted in the older age group compared to the youngergroups, data were combined into two groups: primary(age groups 6-7 years and 9-10 years; elementary grades1-6) and post-primary (12-13 years and 15-16 years; highschool grades 7-11). Mann-Whitney analyses revealeda statistically significant better stereoacuity score for the

post-primary group for the Frisby but not for the TNO(Mann-Whitney; Frisby crossed: z 5 4.67, P \ 0.0001;Frisby uncrossed: z 5 4.67, P\ 0.0001; TNO: z 5 1.35,

Journal of AAPOS

FIG 2. A, Distribution of Frisby (crossed) stereoacuity scores forprimary (black bars) and post-primary (grey bars) school-aged partic-ipants. B, Distribution of Frisby (uncrossed) stereoacuity scoresfor primary (black bars) and post-primary (grey bars) school-agedparticipants.

FIG 3. Distribution of TNO stereoacuity scores for primary (blackbars) and post-primary (grey bars) school-aged participants.

Volume 17 Number 6 / December 2013 Anketell, Saunders, and Little 585

P 5 0.18). Distribution of stereoacuity scores for bothstereotests are shown in Figures 2 and 3.Frisby crossed stereoacuity scores were statistically

significantly higher than uncrossed scores (z 5 �3.84,P\ 0.001). However, post hoc analyses reveal this differ-ence to be significant only in the primary group (primary:z 5 �3.52, P 5 0.0004; post primary: z 5 �1.82, P 50.07). Table 2 provides normative stereoacuity scores forthe Frisby stereotest, indicating the stereoacuity achievedby 95% participants. Median, interquartile range, and95th centile are also provided for the TNO stereotest forthis study population.Analyses of the Frisby and TNO stereoacuity scores

reveals a statistically significant difference between Frisbyand TNO stereoacuity scores for both crossed anduncrossed disparities (Frisby [crossed]: z 5 �11.34, P \0.0001; Frisby [uncrossed]: z5�10.67, P\0.0001). Com-parison of individual Frisby and TNO stereoacuity scoresby Spearman rank correlation identified a weak correla-

Journal of AAPOS

tion between the Frisby and TNO stereotests (Frisby[crossed]: r 5 0.21, P \ 0.005; Frisby [uncrossed]: r 50.25, P\ 0.001).

Discussion

Reference data for the Frisby stereotest for school agechildren with good visual acuity and binocular vision arelimited. This study reports median Frisby stereoacuityscores for primary school participants of 20 arcsec (crosseddisparities) and 25 arcsec (uncrossed disparities), with 95%of children achieving scores of at least 85 arcsec (crossed)and 170 arcsec (uncrossed). Post-primary school partici-pants achieved median stereoacuity scores of 10 arcsec(crossed and uncrossed disparities), with 95% scoring atleast 85 arcsec (crossed and uncrossed). These data identifythe range of normal stereoacuity scores that should beanticipated clinically in 95% of typically developing chil-dren both in primary and post-primary age groups andwithin four discrete age groups (Table 2). Broadly, a Frisbystereoacuity score of 85 arcsec or better is within normallimits.

Review of 95th centile outcomes for each age group maysuggest that a worsening of Frisby uncrossed stereoacuityscores between age groups 6-7 years and 9-10 years. How-ever, no statistically significant difference was notedbetween these two groups. The broader range of stereo-acuities achieved in the 9-10 years age group explains thehigher 95th centile for this group.

The Frisby stereoacuity scores recorded in the currentstudy were noted to be better for the post-primary thanfor primary subjects. Stereopsis development has beenreported to develop rapidly in infancy with gradualimprovement throughout childhood but debate remainsaround the age at which adult levels of stereoacuity areachieved.12,21 Although the present study may supportthe suggestion that an improvement of stereoacuity

Table 3. Data for the Frisby and TNO stereotests in children and adults

Study Sample size Age, years Frisby, arcsec TNO, arcsec

Hall6 67 18-24 Mean, 23 Mean, 63Rosner and Clift25 20 27.4 17/20 (85%) scored 30 or better —Heron and colleagues11 369 3-7 Median, 55.5-17.5 Median, 95.5–53

51 Mean, 20 Mean, 12.9 Mean, 88.7Leat and colleagues12 15 2-5 Mean, 1.66 (45.7) —

17 5-8 Mode, 1.3 (20)15 8-20 Mode, 1.3 (20)

Williams and colleagues5 859 7 — Median, 57.7811 9 Median, 40.4796 11 Median, 38.1

Tomac and Altay14 115 3-5 — Median, 60-170Robaei and colleagues15 2353 12.7 — 96.3% “normal” (120) or better stereoacuityCurrent study 97 6-7 and 9-10 Crossed: median, 20 (IQR, 20-40)

Uncrossed: median, 25 (IQR, 20-40)Median, 60 (IQR, 60-120)

89 12-13 and 15-16 Crossed: median, 10 (IQR, 5-20)Uncrossed: median, 10 (IQR, 10-25)

Median, 60 (IQR, 60-120)

Arcsec, seconds of arc; IQR, interquartile range.

586 Anketell, Saunders, and Little Volume 17 Number 6 / December 2013

occurs between primary and post-primary school agegroups, longitudinal data would be needed to investigatethis further.

The Frisby stereoacuity scores of this study populationare comparable to the findings of Heron and colleagues,11

who reported Frisby median stereoacuity scores rangingfrom 55.5 arcsec at age 3 years to 16.5 arcsec at age 7 years.Leat and colleagues12 also report mode log stereoacuityscores of 1.3 (20 arcsec) in a small sample of 32 subjectsaged 5-20 years. The current study results for post-primary participants are comparable to previous studiesthat reported adult Frisby stereoacuity scores rangingbetween 8 arcsec and 23 arcsec. 6,11,22 Table 3 comparesour data to that of previous studies.

Both crossed and uncrossed disparities were measured inthe present study. Clinically a combination of crossed anduncrossed stereoacuity scores may be recorded as thepatient stereoacuity threshold. Few previous studies havereported whether scores were measures of crossed oruncrossed stereopsis or a combination of both. Woo andSillanpaa23 noted that “crossed disparities are easier todetect than uncrossed disparities” and that responses toboth disparities should be assessed. Within the presentstudy it was noted that crossed stereoacuities were signifi-cantly, albeit only slightly, better than uncrossed in theprimary subjects. This finding may support previouswork reporting that sensitivity to crossed disparitiesdevelops before sensitivity to uncrossed disparities.24

However, a study of young adults by Larson22 reportedno significant difference between crossed and uncrossedstereoacuities, in agreement with data from the post-primary subjects in the present study. These findings maysuggest that early supremacy of crossed stereoacuity islost later in childhood as adult levels are achieved. Thesedifferences between crossed and uncrossed stereoacuitiesin younger children underscore the importance of exam-iners recording which type of stereoacuity has beenassessed when using the Frisby stereotest.

A median TNO stereoacuity score of 60 arcsec wasrecorded for both primary and post-primary subjects inthe current study, with 95% in both groups scoring 480arcsec or better. These stereoacuity scores are comparableto previously published reference data for the TNOstereotest.5,15 Williams and colleagues5 reported medianstereoacuity scores from a longitudinal study in whichTNO was assessed at ages 7, 9, and 11 years. Stereoacuitywas seen to improve from 57.7 arcsec to 38.1 arcsec duringthis timeframe. Robaei and colleagues15 reported that 96%of children (mean age, 12.7 years) scored normal stereoacu-ity (defined as 120 arcsec) or better. In the present study,if the classification of $120 arcsec was used, 83.9% ofsubjects would have been defined as having normal stereo-acuity.

The failure rate of 3.8% in the present study is similar tothat reported by Williams and colleagues,5 which rangedfrom 2.1% to 3.3%. However, if all participants scoring.120 arcsec are included, then 16.1% of the subjects inthe current study would be classed as having reducedstereoacuity. This is considerably higher than the 3.7%reported by Robaei and colleagues15 and the 11.9% and9.1% for age 9 and 11 years fromWilliams and colleagues5

but substantially lower than the 41% noted by Williamsand colleagues5 for children aged 7 years. Robaei and col-leagues15 propose various causes for reduced stereoacuity,including manifest strabismus or amblyopia. However,participants with strabismus and/or an interocular differ-ence in distance visual acuity of $0.2 logMAR and/oranisometropia .1 D were excluded from the present dataset. In an effort to explain why seven children in the presentstudy failed the TNO stereotest, we reviewed individualresults and identified 3 participants with an interoculardifference of near visual acuity of .0.1 logMAR. This dif-ference may lead to a defocused retinal image reducingbinocular potential however; Robaei and colleagues15 sug-gests that a difference in near visual acuity has a limitedeffect on stereoacuity scores.

Journal of AAPOS

Volume 17 Number 6 / December 2013 Anketell, Saunders, and Little 587

Participants in both age groups recorded significantlybetter stereoacuity scores with the Frisby stereotest thantheTNO stereotest. Previous studies have described corre-lations in stereoacuity scores between the Frisby and TNOstereotests ranging from 0.17 to 0.73,6,11,13,25 with thehighest correlation found in an adult study.25 In the currentstudy statistical analyses revealed a positive but weak corre-lation between the Frisby and TNO stereotests, indicatingthat it is notmerely due to a linearly related higher score forTNO. Poor correlation between the two may be due to anumber of factors, including the different scoring criteriafor each test. The TNO was more conservatively scored,with both targets needing to be successfully identified ateach level, whereas only two out of three plates wererequired to be achieved for the Frisby stereotest. Further-more, there were many more possible values for Frisbystereoacuity scores compared to the TNO.Differences in stereoacuity scores between real-depth

and randot stereotests have been well documented.6,13,16

When using the Frisby stereotest monocular cues canoccur if participants move their head and this couldenable participants to achieve better stereoacuityscores.13,6 To avoid this, the test was held directly beforethe visual axis and head movement was discouraged.Simons13 and Hall6 also propose that the images in real-depth stereotests, such as the Frisby, are easier to perceive,whereas the TNO stereotest requires the participant tohave good form discrimination along with good stereoacu-ity to achieve high levels of stereoacuity. Differences in in-terocular contrast induced by the red/green spectacles hasalso been proposed as a cause of degradation or enhance-ment of stereoacuity scores especially in children withamblyopia.26 It has also been suggested that polarizinglenses may dissociate intermittent strabismus or poorlycontrolled phorias, reducing binocular interactions andimpairing stereoacuity.6,16 Finally, it has also beenproposed that the two tests are measuring differentaspects of stereoacuity therefore direct comparison of thescores are inappropriate. Regardless of the cause, poorcorrelation between the scores obtained by the stereotestshas clinical implications. Clinicians should not use thetests interchangeably when monitoring stereoacuity overa series of visits or during a treatment intervention.

References

1. Wallace DK, Lazar EL, Melia M, et al. Pediatric Eye Disease Inves-tigator Group. Stereoacuity in children with anisometropic ambly-opia. J AAPOS 2011;15:455-61.

2. Lee SY, Isenberg SJ. The relationship between stereopsis and visualacuity after occlusion therapy for amblyopia. Ophthalmology 2003;110:2088-92.

3. Walraven J, Janzen P. TNO stereopsis as an aid to the prevention ofamblyopia. Ophthalmic Physiol Opt 1993;13:350-56.

Journal of AAPOS

4. Shipmann S, Cohen KR. Relationship of heterophoria to stereopsis.Arch Ophthalmol 1983;101:609-10.

5. Williams S, Simpson A, Silva PA. Stereoacuity levels and vision prob-lems in children from 7 to 11 years. Ophthalmic Physiol Opt 1988;8:386-9.

6. Hall C. The relationship between clinical stereotests. OphthalmicPhysiol Opt 1982;2:135-43.

7. Saunders KJ, Wodhouse JM, Westall CA. The modified Frisby ster-eotest. J Pediatr Ophthalmol Strabismus 1996;33:323-7.

8. Manny RE, Martinez AT, Fern KD. Testing stereopsis in the pre-school child: is it clinically useful? J Pediatr Ophthalmol Strabismus1991;28:223-31.

9. Cooper J, Feldman J. Assessing the Firsby Stereo Test under monoc-ular viewing conditions. J Am Optom Assoc 1979;50:807-9.

10. Lam SR, LaRoche GR, De Becker I, Macpherson H. The rangeand variability of ophthalmological parameters in normal childrenaged 4½ and 5½ years. J Pediatr Ophthalmol Strabismus 1996;33:251-6.

11. Heron G, Dholakia S, Collins DE, McLaughlan H. Stereoscopicthreshold in children and adults. Am J Optom Physiol Opt 1985;62:505-15.

12. Leat SJ, St. Pierre J, Hassa-Abadi S, Faubert J. The moving DynamicRandom Dot Stereosize test: development, age norms, and compari-son with the Frisby, Randot, and Stereo Smile test. J Pediatr Ophthal-mol Strabismus 2001;38:284-94.

13. Simons K. Stereoacuity norms in young children. Arch Ophthalmol1981;99:439-45.

14. Tomac S, Altay Y. Near stereoacuity: development in preschool chil-dren; normative values and screening for binocular vision abnormal-ities; a study of 115 children. Binocular Vis Strabismus Q 2000;15(3):221-8.

15. Robaei D, Huynh SC, Kefley A, Gole GA, Mitchell P. Stereoacuityand ocular associations at age 12 years: findings from a population-based study. J AAPOS 2007;11:356-61.

16. Leske DA, Birch EE, Holmes JM. Real depth ersus Randot stereot-ests. Am J Ophthalmol 2006;142:699-701.

17. Birch E, Williams C, Hunter J, Lapa MC. ALSPAC “Children inFocus” Study Team. Random dot stereoacuity of preschool children.J Pediatr Ophthalmol Strabismus 1997;34:217-22.

18. Broadbent H, Westall C. An evaluation of techniques for measuringstereopsis in infants and young children. Ophthal Physiol Opt 1990;10:3-7.

19. Anon. TNO test for stereoscopic vision. 17th ed. Nieuwegein, TheNetherlands: Lam�eris Ootech B.V; 1972.

20. O’Donoghue L, Saunders KJ, McCelland JF, et al. Sampling andmeasurement methods for a study of childhood refractive error in aUnited Kingdom population. Br J Ophthalmol 2010;94:1150-54.

21. Fawcett SL,Wong YZ, Birch EE. The critical period of susceptibilityof human stereopsis. Invest Ophthalmol Vis Sci 2005;46:521-5.

22. Larson W. An investigation of the difference in stereoacuity betweencrossed and uncrossed disparities using Frisby andTNO tests. OptomVis Sci 1990;67:157-61.

23. Woo GC, Sillanpaa V. Absolute stereoscopic thresholds as measuredby crossed and uncrossed disparities. Am J Optom Physiol Opt 1979;56:350-55.

24. Birch EE, Gwiazda J, Held R. Stereoacuity development for crossedand uncrossed disparities in human infants. Vision Res 1982;22:507-13.

25. Rosner J, Clift GD. The validity of the Frisby stereotest as a measureof precise stereoacuity. J Am Optom Assoc 1984;55:505-6.

26. Simons K, Elhatton K. Artifacts in fusion and stereopsis testing basedof red/green dichoptic image separation. J Pediatr Ophthalmol Stra-bismus 1994;31:290-97.