TOPICAL REVIEW

Standardization of Reading Charts: A Review of

Recent Developments
Wolfgang Radner, MD1,2*
Downloaded from http://journals.lww.com/optvissci by 8rZ/p9VvOfK8tPKKNf9xan/i2kTS5fE+sHufNNcGhsif7IoSP4yyj+1MJUa/qc9CcvpPRF4cFYu4BxvJBuBywczo8CQJwfCVV9A6Za3CoJPqif0AcBADBY1C56jjQ5MS on 10/08/2019

ABSTRACT: This review gives an overview of the current status of standardization by statistical evaluation of reading
charts. First begun only 20 years ago, the statistical evaluation of reading charts now reflects an increasing clinical
and scientific interest in standardized, comparable, and reproducible reading charts.
For clinical or research purposes in human subjects, standardization of psychophysical tests and their test items by
statistical evaluation is mandatory because it provides experimental control. Initial attempts at reading chart stan-
dardization were made by characterizing the test items, either in terms of a selection of unrelated words or in terms
of sentences representing a constant number of characters, including spaces. As initiated by the RADNER Reading
Charts, standardization of reading charts (and test items) by statistical evaluation has gained increasing clinical
and scientific interest in the last two decades and has later also been applied to some of the other modern reading
charts. A literature search was performed with respect to reading charts that (a) have been produced in accordance
with the recommendations of the International Council of Ophthalmology (geometrical print size progression), (b)
have been statistically analyzed, and/or (c) use clearly characterized test items (conceptually and statistically). These
reading charts are as follows: the Bailey-Lovie Word Reading Charts, the Colenbrander Cards, the RADNER Reading
Charts, the MNREAD Acuity Charts, the Smith-Kettlewell Reading Test (SKread Test), the C-Read Charts, and the Author Affiliations:
1
Balsam Alabdulkader-Leat (BAL) Chart. The test items of these charts have been characterized either empirically Austrian Academy of Ophthalmology,
or by statistical analysis and selection. The extent of the statistical evaluation of the reading charts varies. Despite Vienna, Austria
2
their different methodological approaches, these reading charts represent an advancement that has made possible Department of Ophthalmology,
the useful comparison and reproducible evaluation of near visual performance. University Hospital St. Pölten, Karl
Landsteiner University for Health
Optom Vis Sci 2019;96:768–779. doi:10.1097/OPX.0000000000001436 Sciences, St. Pölten, Austria
Copyright © 2019 American Academy of Optometry *wolfgang.radner@chello.at

Standardization by statistical evaluation is mandatory for psy-
chophysical tests and their test items when the tests are to be used
for clinical or research purposes in human subjects. This testing is
a crucial ingredient in the interpretation, comparability, reproduc-
ibility, validity, and reliability of such tests.1–3 In the case of tests
of the intelligence quotient, it is commonly taken for granted that
the validity, reliability, and reproducibility of the tests are known.
We expect that these parameters have been investigated and that
the variability of the results is within an acceptable range. Stan-
dardization by statistical evaluation can thus be considered as a
tool that provides experimental control, a basic requirement when
cognition, perception, or recognition is being investigated.2,3
These expectations are also true for psychophysical tests such
as visual acuity and reading charts. For example, if the mean reading
speed of a patient is found to be 200 words per minute at a print size
of 0.5 logRAD and 160 words per minute at 0.4 logRAD, but the
consistency of the test sentences/test items is not known, it is not
possible to determine whether the sentence characteristics or the
print sizes are responsible for the difference in reading speed. In
addition, without information about validity, it also remains unclear
which of the two results is more representative of the patient's
reading performance.
It is therefore evident that, irrespective of the structural specifica-
tions of the test items, only standardization by statistical evaluation
can provide appropriate comparability and investigational control.3
Furthermore, when such test items are used in a reading chart,
statistical investigations of the reliability, reproducibility, and validity
of the reading chart are also useful.
Unfortunately, statistical evaluation has not been congruently
established for reading charts, as it has for distance acuity charts.4
With the exception of some earlier improvements,4–7 it was not
until 1983 when the International Council of Ophthalmology spec-
ified how comparable geometrically (logarithmically) progressing
print sizes of continuous text reading charts could be achieved.8
Nevertheless, a number of uncontrolled remakes of historic reading
charts, such as Jaeger, Nieden, and Parinaud charts, and reading
charts printed as advertisements were put into use (and are still
in use)4; this reworking has occurred in the absence of any stan-
dards. The print sizes of such reading charts do not progress geo-
metrically (logarithmically), and the remakes often do not even
use the same print sizes as their original archetypes (Colenbrander
A, Runge P. IOVS 2007;48:ARVO E-Abstract 3563),4,9–13 making
it likely that use of the original developer's name has not been
authorized.4,13 This lack of standardization has resulted in (a) a
confusing variety of reading acuity measures, (b) an incomparability
of results obtained with different reading charts, (c) weak information
about patients' near visual performance, and, most unfortunately, (d)
a clinically and scientifically underestimated measurement of near
visual acuity.4
The premise for the development of the RADNER Reading Charts
was to establish for reading charts a standardization at a level
equivalent to that for psychological test systems, with the goal of

www.optvissci.com Optom Vis Sci 2019; Vol 96(10) 768

Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.

 Reading Chart Standardization — Radner

achieving a comprehensive comparability between reading
charts.4,14–19 Later, other reading chart systems followed this ex-
ample to various extents.4,20–23 This article is intended to outline
how far standardization by statistical evaluation has permeated this
field and to spark interest in attempting to achieve international
comparability through the statistically evaluated of reading charts.4
STANDARDIZATION BY STATISTICAL EVALUATION
AND/OR SPECIFICATION OF TEST ITEMS
Standardization (and selection) of test items by statistical eval-
uation was introduced for reading charts in the late 1990s for the
sentence optotypes of the RADNER Reading Charts.14–16 This stan-
dardization included analyses of reliability (consistency across test
items) and validity (degree to which a test item measures what it is
supposed to measure). To reduce the variation and increase the
consistency between the sentence optotypes, the analysis further
included selection of eligible sentences on the basis of defined up-
per and lower limits for a narrow range of mean values for reading
speed and error count intervals. For other reading charts, uniformity
of the test items was achieved either by defining the number and
length of the unrelated words presented per line or by specifying a
constant number of characters (including spaces) and by the difficulty
level of the words used (Bailey-Lovie Word Reading Charts,
Colenbrander Cards, Minnesota Low-Vision Reading Test (MNREAD)
Charts, Smith-Kettlewell Reading Test [SKread Test]).24–26 Later, Ar-
abic and Chinese reading charts varied aspects of these principles or
modified them according to the specific characteristics of their re-
spective languages.20–23
To be consistent with methodological requirements for the standard-
ization of test items and test setup as postulated for psychology,1–3 the
term standardization is used here only when proper statistical anal-
yses have been performed.1–4 Other specifications of test items or
test setups are mentioned here with regard to how they were de-
fined, specified, or characterized according to empirically chosen
features (e.g., specification by the number of characters, or use
of unrelated words).24–26
Statistical Evaluation and Conceptual Standardization
of Test Items (Concept of Sentence Optotypes)
Test Items of the RADNER Reading Charts
Based on the idea of adopting a statistical approach to test stan-
dardization (as used in psychology) for reading charts, a collabora-
tive research group was founded in 1995, involving a psychologist,
a linguist, and an ophthalmologist.14,15 The initial target was to
create “sentence optotypes” that were consistent with each other,
keeping the geometric proportions constant to allow accurate mea-
surement of reading parameters.14,15
A series of test sentences was originally generated in German
and subsequently in 11 more languages (main clause followed by
a relative clause), all of which had to be as comparable as possible
in terms of the number of words (14 words), word length, number of
syllables per word, position of words, number of characters, lexical
difficulty, and linguistic aspects such as grammar and syntax.14,15
These sentence optotypes of 3 lines and 14 words incorporated 82
to 84 characters, including spaces (27 to 29 characters per line)
and 22 to 24 syllables. In addition, the position and length of
the words were defined by specific rules.14,15 Using a group of
198 volunteers, the most statistically equivalent sentence
optotypes were selected by measuring (a) Cronbach's α, (b) the
corrected item total correlation, (c) α if item deleted, and defining
(d) narrow intervals with respect to reading speed (Fig. 1) and diffi-
culty (overall mean ± 0.25  standard deviation; Fig. 1).14,15,17 Ac-
cordingly, until now 62 sentence optotypes were statistically
selected for the German Reading Charts.14,15,17
The Cronbach's α and the corrected item total correlation were
high and well above statistically required limits14,15,17; α if item
deleted never exceeded the Cronbach's α coefficient (Fig. 2). The
reading speed correlated well with that obtained for the standardized
long paragraphs 3, 4, and 5 of the Zuercher Reading Test (standardized

FIGURE 1. Example showing the mean reading speed and SD (standard deviation) of 24 sentence optotypes selected for the RADNER Reading Charts
and the reading speed interval that is used for sentence selection (reading speed interval = overall mean (dashed line)  ± 0.25  SD (solid red lines).
Sentences are eligible for the reading chart when their mean is within the interval for reading speed and reading errors. Reproduced with permission from
Radner et al.15

www.optvissci.com Optom Vis Sci 2019; Vol 96(10) 769

Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.

Reading Chart Standardization — Radner

FIGURE 2. Plot of statistical parameters for the RADNER Reading Charts showing comparative Cronbach's α values for the preselected 34 sentences
(solid line) and the 28 finally selected sentences (dashed line), α if item deleted for the 28 selected sentences, and corrected item total correlation
for the sentence optotypes of the English version. The Cronbach's α and the corrected item total correlation were well above statistically required limits;
α if item deleted never exceeded the Cronbach's α coefficient. Reproduced with permission from Radner et al.17

for fourth to sixth grade), indicating the high validity of these test
items.15 In accordance with the concept of sentence optotypes,
optimized sentences that were constructed so that words having
the same number of characters and syllables were placed in the
same positions were later developed. Because all the statistical re-
sults obtained with these optimized sentences were similar to
those of their corresponding original sentences, the best possible
reliability had apparently already been achieved with the original
sentence optotypes.27 The concept of sentence optotypes has
now been applied to 12 different languages, with language-specific
modifications. A total of 1323 volunteers have been tested to
standardize sentence optotypes in 12 languages: German, Spanish,
English, French, Dutch, Italian, Swedish, Danish, Portuguese,
Turkish, Hungarian, and Romanian.14–17,19,27–34
Specification of Unrelated Words (Bailey-Lovie Word
Reading Charts)
Bailey and Lovie24 designed a word-reading chart with a logarithmic
size progression that used unrelated words. They decided to use
4-, 7-, and 10-letter words at each size level, based on the observation
that, in patients with age-related macular degeneration, the word
length can affect the readability (some patients prefer longer words,
others shorter ones). The words and word order were selected with
the intention of having the first letters of the words evenly distributed
over the whole alphabet. The frequency of word use also became a
selection criterion, and care was taken to avoid obvious syntactic as-
sociations between adjacent words.24,35
Specification of the Number of Characters and Spaces
(Colenbrander Cards, MNREAD Charts, SKread Test)
Test Items of the Colenbrander Continuous Text Near
Vision Cards
All of the test items (sentences) used in the Colenbrander
Continuous Text Near Vision Cards (Precision Vision, Woodstock,
IL) have 44 characters, including spaces. The sentences have
9 to 11 words. Difficulty of the sentences is controlled by limit-
ing the sentence length and by rejecting long words (grade level,
about fourth grade). For decimal acuities from 0.063 to 0.1, one
sentence is presented per print size, and for 0.12 and smaller,
two sentences are presented. However, the test sentences of the

www.optvissci.com Optom Vis Sci 2019; Vol 96(10) 770

Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.


Colenbrander Cards have not been developed and specified for
reading speed measurements.
Test Items of the MNREAD Charts
For the initial version of the MNREAD Charts that was published
in 1993, the sentences of the computer-based MNREAD test were
used.25,36 These sentences incorporated 4 lines and 52 characters
including spaces (13 characters per line).36 Later, a new pool of
sentences has been developed for the current version of the
MNREAD Charts (three lines; 60 characters including space). These
sentences also incorporated the structural and linguistic criteria of
the MNREAD test and a restricted vocabulary; the majority of the
words are among the 1000 most frequent words in third-grade
school books.37 The sentences are characterized by incorporating
the concept of standard-length word, as postulated by Carver38,39
for English text passages. Carver38,39 found that the average word
length as calculated from the number of characters, including
spaces (character-spaces), divided by the actual number of words
increases with the difficulty of text passages. For estimating read-
ing speed obtained with English text passages, he postulated that
using six character spaces as the definition of a standard-length
word made it possible to reduce the variation in reading speed be-
tween text passages with different difficulty grades.38,39 Accord-
ingly, the sentences of the MNREAD Charts have 60 characters,
including spaces (character spaces), with an implied period at
the end of the sentence. Thus, 60 character spaces equal 10
standard-length words of six character spaces (the actual number
and length of words vary between sentences). This turned out to
be convenient for scoring reading errors and reading speed, be-
cause it helped to reduce the variation in scoring that occurred as
the result of the different word lengths and word number found in
different sentences.4,25,37
Carver39 found that a mean word length of six character spaces
represented difficulty grade 12.8 (of 17 grades). As reported by
Carver,39 the reading speed of college students measured by the
actual word count decreased from about 315 to 200 words per
minute as the difficulty increased from difficulty grade 2 to about
grade 17; in contrast, when reading speed was measured for
standard-length words, the reading rates became less variable at
about 250 to 260 words per minute, as the difficulty increased
from about grade 5 to about grade 14. Thus, Carver39 concluded
that “when reading rate is measured in actual words per minute,
such estimates will tend to overestimate rate when the reading ma-
terial is grade 12 difficulty or less and will tend to underestimate
rate when the reading material is grade 13 difficulty or greater.”
The MNREAD Charts are available in several languages. However,
Carver's concept has only been verified for English.21
Test Items of the SKread Test
The SKread Test (Precision Vision, Woodstock, IL) was devel-
oped to assess the reading performance of low-vision patients
and simultaneously allow estimation of the location of scotomas
in the macula.4,26 The test paragraphs consist of sequences of ran-
dom (unrelated) words and letters and have been specified according
to the MNREAD format (60 characters, including spaces). Each test
paragraph contains 6 single letters and 10 unrelated, randomly
chosen words. By analogy to the concept of sentence optotypes
of the RADNER Charts, the number of words is the same in all par-
agraphs. In addition, the number of words with two, three, four,
five, and six letters is equal in all paragraphs. Words that can stand
www.optvissci.com Optom Vis Sci 2019; Vol 96(10)
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
Reading Chart Standardization — Radner
alone with letters missing from the beginning or the end of the word
were especially included. This test principle was chosen because
the authors wanted reading performance to depend upon word
and letter recognition alone and wanted to exclude linguistic as-
pects such as grammar and syntax.26
Test Items for Chinese and Arabic Reading Charts
Chinese Charts by Cheung et al. and C-Read Charts
There are two Chinese reading charts.22,23 The one developed
by Cheung et al.22 uses 12 traditional Chinese characters per line,
whereas the C-Read Chart uses 12 simplified Chinese characters.23
Because simplified and traditional Chinese characters are read with
different fluency depending on a person's place of origin, and be-
cause traditional Chinese characters have significantly higher spa-
tial complexity than their simplified counterparts, Han et al.23 have
suggested that the simplified and traditional Chinese reading
charts should be considered two different instruments for two dif-
ferent populations.
The reading chart version developed by Cheung et al.22 were de-
veloped for children. The test sentences to be used were selected in
the following way: 17 dyslexic and 9 nondyslexic children were re-
cruited for empirical testing. Using a concept similar to that of the
sentence optotypes of the RADNER Reading Charts,14–17 an inter-
val for reading speed and errors was used for the final selection of
the test items. The sentences with the highest and lowest mean
reading time, those with the largest intersubject variance in read-
ing time, and those that induced the highest number of reading er-
rors were excluded. In all, 51 sentences with a standard deviation
in reading time of less than or equal to 1 second and minimal read-
ing errors of less than or equal to one word for nondyslexic children
and less than or equal to two words for dyslexic children were se-
lected to construct three versions of a Chinese reading chart
for children.22
For the C-Read Charts,23 105 simple declarative sentences
were chosen from first- to third-grade textbooks of Chinese elementary
schools. These sentences were reduced by pre-selection with re-
gard to the characteristics of Chinese characters. Again following
the concept of sentence optotypes of the RADNER Reading
Charts,14–17 the remaining 67 sentences were then statistically in-
vestigated for reading speed, reading error, and reading fluency in
20 Chinese college students. Sentences were not selected on the
basis of time or speed intervals, but instead the subjects subjec-
tively rated the suitability of each sentence for reading assessment
on a 1-point (least suitable) to 5-point (most suitable) scale. The
sentences were then further simplified by eliminating those that
contained characters that could easily be confused with others,
and sentences with the longest and shortest reading times and/or
with the lowest suitability scores were removed until 48 sentences
remained. After statistical and empirical specification, fine adjust-
ments were made by replacing less frequently used characters and
by limiting the number of very simple characters (three strokes or
fewer) to no more than three.23
Arabic — Balsam Alabdulkader-Leat (BAL) Chart
Like those of the RADNER Reading Charts,14–17 the Arabic test
sentences of the BAL Chart were statistically analyzed, and the
reliability was investigated by calculating a Cronbach's α quo-
tient in a group of 69 persons (53 adults, 16 children).20,21
Mean reading speeds and reading errors were calculated, and
771
Reading Chart Standardization — Radner
outliers were excluded. Cronbach's α was similar to that ob-
tained from the sentence optotypes of the RADNER Reading
charts. The paired sentence design and the specification of the
sentences were established in accordance with the Colenbrander
charts. The number of characters, including spaces, for each pair
of sentences was 102 characters; the number of words in each pair
ranged from 16 to 22, and no word had more than 10 letters.20
Reading speed was calculated either on the basis of correctly read
words or, as in the MNREAD Charts, by using Carver's38,39 concept
of standard-length word (number of characters including spaces, di-
vided by the number of words). The postulation of Carver38,39 that
the standard word length of diverse English paragraphs is six letters
per word was not found to be true for Arabic.20,21 The authors ana-
lyzed articles in Arabic and found that the standard word length
was about 4.7 letters per word; for calculation of reading speed, they
rounded that to five letters per word. As introduced by the RADNER
Reading Charts, one criterion for the final selection from the test
sentences was based upon the calculation of a reading speed interval
around the mean reading speed for the sentences.14–17 This range
was mean ± 1.645  standard deviation for the adults and
mean ± 1.96  standard deviation for the children. These chosen
intervals were wider than that used for the RADNER Reading
Charts (mean ± 0.25  standard deviation).14,15,17 Pairs of
sentences that fell outside these ranges were excluded.
STANDARDIZATION OF READING CHARTS BY
STATISTICAL EVALUATION
A principal question is how controlled test conditions and a
predictable accuracy of the results can be achieved with psycho-
physical test systems such as reading charts. Is it enough to just
analyze the coefficient of repeatability, or is it necessary to estab-
lish a statistical test battery, by analogy to psychological tests, to
provide information about further aspects that characterize the
quality of a test?1–4,14–17,19,27 Irrespective of the approach chosen,
more detailed analyses of a test system are not disadvantageous.
Therefore, given that there was no existing example that could have
been adopted for the standardization of reading charts, we established
a statistical test battery for the RADNER Reading Charts that in-
cluded test-retest reliability, interchart reliability, variance compo-
nent analysis, and Bland-Altman analyses.16,19 To our knowledge,
this is the first example of analyzing a reading chart in this way.
The rationale behind this combination of statistical methods is
that statistical evaluations of test-retest and interchart reliability
(reproducibility) can only establish whether a particular source of
variation (e.g., different charts, two tests, different patients) has a
significant influence on the variability of the testing procedure.3
The variance component analyses, on the other hand, quantify
the contributions of these sources to the total variability.3,16 The
advantage of including a variance component analysis is that it
allows us to identify the sources of variability for all measured
variables. In addition, because correlation coefficients indicate
the strength of a relationship between two variables, but not the
agreement between them, Bland-Altman analyses were chosen to
analyze the agreement between two different series of measure-
ments40; they identify systematic differences between measure-
ments and uncover outliers.
This battery of statistical methods provides a detailed view of the ca-
pability, reliability, and accuracy of a reading chart system and has
also been adopted in part for other reading charts.4,20–23,26,41–43
www.optvissci.com Optom Vis Sci 2019; Vol 96(10)
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
RADNER Reading Charts
The RADNER Reading Charts were investigated in terms of test-
retest reliability in a clinical setting (test interval, 3 to 4 weeks),
interchart reliability, reproducibility, repeatability, variance com-
ponent analysis, and Bland-Altman analyses.16,19 Subjects with
acuities between normal and low vision were grouped according
to their distance visual acuity (group 1, 0.1 logMAR or better; group
2, 0.12 to 0.4 logMAR; group 3, 0.42 to 0.8 logMAR). The param-
eters investigated were as follows (orthogonal Latin square design):
(a) reading acuity, (b) reading acuity score, (c) mean reading
speed, (d) maximum reading speed, (e) critical print size, and (f )
logMAR/logRAD ratio (ratio between distance acuity and reading
acuity in percent). Good reproducibility and repeatability were
demonstrated for all reading parameters and visual acuity
groups.16,19 The Bland-Altman plots showed a high degree of
agreement between the two test sessions and among the three
reading charts (Fig. 3). In addition, a variance component analysis
was performed.16 Taken together, these analyses revealed that the
individuals (patients) were predominantly responsible for the vari-
ability of the results. The testing procedure had only a minor influ-
ence on the whole variance, indicating that the test is highly
reproducible.4,16,19 Comprehensive statistical analysis, as demon-
strated by this battery of tests, gave extensive insight into the qual-
ity of the reading charts.
MNREAD Charts
Virgili et al.41 investigated a group of children using the Italian
version of the MNREAD Charts and showed good coefficients of re-
peatability (test interval, the same day) for reading acuity, reading
speed, and critical print size.4 Neither variance component analysis
nor Bland-Altman plots were conducted. In 2006, Subramanian and
Pardhan42,43 reported a good coefficient of repeatability between
the two MNREAD Charts in 30 young adults and later in visually im-
paired patients (test interval, the same day).4 As introduced for read-
ing charts by the RADNER Reading Charts,16 Bland-Altman plots
were included, but variance component analysis was not performed.
Patel et al.44 later reported weaker coefficients of repeatability
(Fig. 4) in age-related macular degeneration patients investigated
with the MNREAD Charts than did Subramanian and Pardhan.43
They concluded that patient- and examiner-related aspects of
variability might be responsible for these differences.44 Also,
practice effects resulting from a much shorter test-retest interval
could not be fully excluded (test interval, [a] the same day in the
study of Subramanian and Pardhan43 and [b] 6 weeks in the
study of Patel et al.44).
SKread Test
Interchart reliability was analyzed in 30 subjects (healthy eyes)
who read with their habitual glasses five paragraphs of SKread
Chart 1 and then of SKread Chart 2 (test interval, a few minutes).26
The print sizes read were the same in both readings. Reading per-
formance was expressed as mean reading time for each set of five
paragraphs and not as reading speed in words per minute, because
the authors considered it inappropriate to treat the single letters in
the SKread as words. Both the correlation coefficient and the coef-
ficient of determination were good. The coefficient of repeatability
between the two reading charts was 9% of the mean reading time
(9.458 seconds). Test-retest reliability was not performed for
each chart.
772
 Reading Chart Standardization — Radner

FIGURE 3. RADNER Reading Charts: intersession and interchart repeatability (test-retest interval: 3 to 4 weeks). Bland-Altman plots exemplifying test-re-
test analyses of reading acuity and maximum reading speed (fastest sentences read) obtained from patients with macular diseases of three visual acuity
groups between –0.1 logMAR and 0.8 logMAR; mean reading acuity of 0.30 ± 0.3 logRAD (range, −0.1 to 0.8 logRAD; n = 36).16 (A, B) Test-retest
analyses comparing results obtained with chart 1 in test sessions 1 and 2 for (A) reading acuity and (B) reading speed. (C, D) Bland-Altman plots com-
paring (C) reading acuity and (D) reading speed obtained in test session 1 for chart 1 and test session 1 for chart 2. (E) Reading acuity and (F) reading
speed analysis results obtained in test session 1 for chart 2 compared with results in test session 2 for chart 3. A and B reproduced with permission from
Stifter et al.16 C, D, E, and F plotted from data published in Stifter et al.16

www.optvissci.com Optom Vis Sci 2019; Vol 96(10) 773

Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.

Reading Chart Standardization — Radner

FIGURE 4. MNREAD Charts: intersession repeatability (test-retest interval: 6 weeks). Bland-Altman plots exemplifying test-retest analyses of reading
acuity and maximum reading speed obtained from age-related macular degeneration patients (mean reading acuity, 0.22 ± 0.19 logMAR; n = 59).44
A, Test-retest analyses comparing results obtained in test sessions 1 and 2 for reading acuity. (B, C) Test-retest analyses comparing results obtained
in test sessions 1 and 2 for two different definitions of maximum reading speed: (B) means of the three sentences read the fastest and (C) the speed
of the sentence read the fastest. Bland-Altman plot of differences against the mean for reading acuity and maximum reading speed. Three horizontal
dotted lines: the upper limit of 95% confidence interval and the mean and lower limit of 95% of the difference (top to bottom).44 Reproduced with permission
from Patel et al.44

www.optvissci.com Optom Vis Sci 2019; Vol 96(10) 774

Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.

 Reading Chart Standardization — Radner

C-Read Arabic — BAL Chart

The C-Read Charts were standardized in 30 native Chinese-
speaking college students by comparing the three charts to each
other.23 By analogy to the testing of the RADNER Reading
Charts,16 (a) interchart correlations were calculated for the reading
parameters (reading acuity, critical print size, maximum reading
speed); (b) agreement between the three charts was shown by
Bland-Altman plots (Fig. 5)40; and (c) a variance component anal-
ysis was performed. The interchart reliability and agreement among
the charts were good. The variance component analyses revealed
that the probands were predominantly responsible for the variabil-
ity of the results, indicating that the test is reproducible.
Because the design of the Arabic typeface is different from Latin
typefaces, the Arabic letter heights of the BAL Charts were calibrated
to logMAR based upon the reading acuity results obtained from par-
ticipants who also read the MNREAD and the RADNER Reading
Charts.20 Statistical evaluation was performed by comparing the
three BAL Charts with each other, and interchart reliability and
Bland-Altman analyses were performed (Fig. 6).20 In addition,
the reading speed results were compared with those for the Arabic
version of the International Reading Speed Texts (IReST)45 and re-
vealed good correlations. Interchart reliability revealed statistical
differences between BAL Chart 1 and the other two charts in terms

FIGURE 5. C-Read Charts (test-retest interval, same day): Bland-Altman plots between different versions of the C-Read. The difference for reading acuity
and reading speed between each version is plotted against the mean. A, Chart A versus chart B for reading acuity and (B) for reading speed. C, Chart B
versus chart C for reading acuity and (D) for reading speed. Red solid lines represent the means of difference measured with a pair of charts. The pair of
black dashed lines represents the 95% limits of agreement. The pair of gray dotted lines around the black dashed line represents the 95% confidential
interval for the limit of agreement.23 Reproduced with permission from Han et al.23

www.optvissci.com Optom Vis Sci 2019; Vol 96(10) 775

Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.

Reading Chart Standardization — Radner

FIGURE 6. Balsam Alabdulkader-Leat Charts (BAL Charts; test-retest interval, same day). Bland-Altman plots between different versions of the BAL
Chart. The speed (SLWPM = log Standard Length Words per Minute) is plotted against the mean. A, Agreement between BAL1 and BAL2 for reading
acuity and (B) for reading speed. C, Agreement between BAL1 and BAL3 for reading acuity and (B) for reading speed. Dashed lines are 95% con-
fidence intervals. The solid blue line represents the mean.20 Reproduced with permission from Alabdulkader et al.2

of reading speed. There were statistically significant differences
among all three charts in terms of reading acuity, but because the
difference was less than half a line (<0.05 logMAR) in each case,
it was described as not being clinically significant.46 Although such
statistical differences should be kept in mind for research purposes,
Bland-Altman plots and the coefficients of agreement showed a
good agreement among the three charts.20
DISCUSSION
This review article is meant to show how far reading chart stan-
dardization has evolved in the last two decades and to point out
how it has fortunately become increasingly seen as a mandatory re-
quirement for reading charts used in clinical practice and research.
Although the definition of reading parameters and the selection of
statistical methods vary among the reading charts,14–17,19–27 the
advantage of using statistically evaluated reading charts for clinical
practice is unequivocal, as has been demonstrated in many research
studies.4,18 In the following paragraphs, some of the most important
aspects that can limit the comparability are exemplified.
Stages of Visual Perception
Vision involves at least three stages: (1) the optical stage, which
projects an image on the retina; (2) the receptor stage, which con-
verts that image into neural impulses; and (3) neural processing,
which leads to perception and recognition and ultimately to visually
guided behavior.47
Optotype recognition involves only the retinal area where the
symbol is projected (stage 1). Even for a 20/200 letter (1.0
logMAR), this is equivalent to less than 1°.47 Reading tests require
a larger retinal area (stage 2) and therefore are more relevant to ac-
tivities of daily living.47
All tests require neural processing (stage 3), which starts in the
inner retina and proceeds to the visual cortex and to higher cortical
centers. For reading tests, the amount of cognitive processing is
larger than for letter charts.35,47,48 This difference explains the
well-known fact that letter chart acuity alone is a poor predictor of
overall visual performance and even of reading performance.47,49,50

www.optvissci.com Optom Vis Sci 2019; Vol 96(10) 776

Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.


Reading Speed
Because the test items (sentence optotypes) of the RADNER
Reading Charts are highly consistent with each other and have
the same number of words, reading speed can be directly calculated
from the actual number of words and time needed to read a sentence
(840 divided by the reading time in seconds).14,15,17,28–34 Read-
ing errors are not considered in reading speed calculations because
a stop criterion is used. Thus, the evaluation of reading speed is
easy to perform and has been demonstrated to produce valid read-
ing speed results.4,14–19 In other reading charts systems, reading
speed is calculated based upon the number of words read cor-
rectly20,23,25,26; in this case, the number of words varies between
the sentences (paragraphs), making reading speed calculations
more complex. Furthermore, when the number of words within a
sentence is calculated using Carver's38,39 principle of standard-
length words (six character spaces), the actual number of words
within a sentence differs from this estimate. Carver38,39 postu-
lated that using standard-length words mathematically reduces
the variability of reading speed results. In addition, the variability
might be further affected when incorrectly read words are excluded
from reading speed calculations and (a) the actual number of
words within a sentence is different from that of standard-length
words or (b) an excluded word is much longer than the standard
word length.
In addition, reading speed has been analyzed either by using the
actual reading speed in words per minute15–17,19 or after logarithmic
conversion to log(words per minute).41–43 However, logarithmic con-
version also reduces the apparent variability and thus makes it even
more difficult to judge or compare the reliability and reproducibility
between different studies. Recently, Calabrèse et al.51 recom-
mended for the MNREAD Charts that reading speed should also be
presented in words per minute rather than in log(words per minute)
when reading speed is normally distributed within a sample. This ad-
vice is consistent with the approach chosen for the RADNER Read-
ing Charts.14–17,19 Because reading speed is commonly normally
distributed within a population, the reading speed should generally
be reported in words per minute and not in log(words per minute).
Accordingly, words per minute is more useful and informative for
practitioners or institutions with legal obligation to make financial
decisions about social support.
Reading Acuity
Although reading acuity is mainly related to print size, the strat-
egies for scoring reading acuity vary slightly. For the RADNER Read-
ing Charts, reading acuity is determined as the smallest print size
read, by using a stop criterion (related to reading time and distortion
of content).14–16 Reading errors can be implemented separately in
the form of a logRAD score by adding an error score, calculated on
the basis of the number of syllables of the incorrectly read words,
to weight different word lengths (the value of one syllable is 0.005
logRAD; longer words with more syllables have a higher impact on
the understandability of a sentence than do short words, and therefore
the score).14 Other reading charts using the principle of standard-
length words include reading errors as part of reading acuity.25,37 In
this case, sentences of 60 characters and spaces equal 10
standard-length words of 6 letters, and every incorrectly read word
has a value of 0.01 logMAR. This correction is independent of word
length and the actual number of words within a sentence. For exam-
ple, the formula recommended for the MNREAD Charts is as follows:
reading acuity = 1.4 − (sentences  0.1) + (errors  0.01). Although
www.optvissci.com Optom Vis Sci 2019; Vol 96(10)
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
Reading Chart Standardization — Radner
both the repeatability and variance component analyses gave good
results,41–44 methods that incorporate reading errors into reading
acuity have limits because a sentence cannot be scored as being
read when more than half of the words have not been read correctly
(i.e., when the content of the sentence is significantly distorted).
Thus, the linear inclusion of errors cannot be completely fulfilled.
Incorporating reading errors may provide additional information
or may represent a refinement of reading acuity measurements;
however, their inclusion does not work like linear Early Treatment
Diabetic Retinopathy Study (ETDRS) letter counts.52
Critical Print Size
Critical print size can be defined in many different
ways. 14,15,41,42,44,53 Some methods are mathematically so-
phisticated and/or based on log(words per minute), for example,
defining critical print size as the smallest print size that yields
90% of the mean reading speed of all the preceding sentences.41
Patel et al.44 defined the critical print size as the smallest print size
that supports a particular reading speed at both 90% and 80% of
the fastest-read three sentences. For the RADNER Reading Charts,
we preferred to use the examiner's decision, an approach that is closer
to clinical practice and defines the critical print size as the smallest
print size a patient can read with normal reading speed before a
notable decrease in reading speed becomes apparent.4,16,19,53
In a study of the age-related course of best-corrected reading acu-
ity and speed,53 we also analyzed the mathematical and clinical
approaches to determining the critical print size. It was particu-
larly interesting that a better correlation with reading acuity and
a higher degree of similarity between the age-related course of
acuity and other reading parameters seemed to indicate that the
conventional decision of an examiner is at least not worse, or
may be even better, than the (post hoc) mathematical approach.53
Nevertheless, in published studies, the critical print size has been
found to give the weakest statistical results with regard to repeatabil-
ity and variance component analyses.16,19,23,41,42,44 It is therefore
likely that critical print size is more of a clinically relevant parameter.
Stop Criteria
A stop criterion of 20 seconds is suggested for the sentences of
the RADNER Reading Charts,4,14,53 corresponding to a reading
speed of about 40 words per minute, which represents the higher
limit of spot reading.54 Additional rules for termination include a
considerably distorted content or when more than 7 of 14 words
per sentence are read incorrectly.4,53 In most of the studies about
reading chart standardization, no stop criterion is given.20,24,41,42
In a study of the C-Read test, the test session was ended when
the subjects reported that they could no longer read any more Chinese
characters.23 For the MNREAD Chart, the recommended instruction
for tested subjects is to read through all sentences until they can no
longer read any words.37
Scoring procedures for reading performance can be divided into
two types: (a) scoring based upon a stop criterion combined with a
line-by-line (or sentence-by-sentence) procedure4,53; in this case,
reading performance is measured at the smallest print size a person
can read within the stop criterion, and particular weight is given to
errors in only the last sentence (irrespective of incorrectly read
words at previous print sizes). (b) Scoring based upon a letter-
by-letter (or word-by-word) procedure21,24,37: in this procedure,
all incorrectly read words from the whole reading chart are included
in the subtraction for determining reading performance. Carkeet's55
mathematical analyses for the ETDRS and Bailey charts have
777
Reading Chart Standardization — Radner

demonstrated possible biases for letter-by-letter scores, which were
explained as a consequence of subjects' correctly guessing letters
that are well below the threshold. Accordingly, Carkeet55 recom-
mended for the ETDRS and Bailey charts to use a line-by-line pro-
cedure and a termination rule of four mistakes (or more) on a line,
because this rule was found to give close-to-optimum slope-
corrected standard deviations.
In addition, particularly in the case of simple sentences, guess-
ing of words during text reading is stimulated by the content of the
test items and contextual predictability. Guessing success is also
related to word length, the difficulty level of the words, the number
of words of a sentence, and reading skills. In addition, the guessing
rate depends on visual acuity and increases with lower visual acu-
ity. Thus, a word-by-word procedure is (a) not comparable with
the letter-by-letter counts used for ETDRS charts52 and is (b) likely
to be affected by many uncontrolled variables. Because parameters
such as reading acuity are meant to express a threshold, that is, the
smallest print size that can be read by a person, it seems reason-
able in terms of comparability to try to achieve such measurement
by using well-defined stop criteria combined with a sentence-by-
sentence procedure.
Stop criteria enhance the risk of losing data. On the other hand,
they provide comparability and experimental control. Therefore,
the criteria have to be defined on the basis of clinical requirements
or particular study designs. Twenty seconds representing about 40
words per minute (spot reading)54 and/or distorted content were
found to be useful stop criteria in most of our studies conducted
with the RADNER Reading Charts4,18,53; using, for example, 80
words per minute (the limit for fluent, sense-capturing reading)54
as a stop criterion would be insufficient, because in normal-
sighted persons, 80 words per minute is close to the reading speed
at print sizes just below the critical print size.4,18,53 With an 80
words per minute limit, therefore, the patient's full visual potential
(i.e., best reading acuity) cannot be shown. However, best-corrected
reading acuity is a result that is as important as best-corrected
distance acuity.
In sum, the use of stop criteria combined with a sentence-
by-sentence procedure provides experimental control and facilitates
the attainment of consistency between different examiners in terms of
decisions concerning the limits of reading acuity and speed.
Notations
In 1874, Snellen and Landolt mentioned in their chapter: “It is
regrettable that for the determination of visual acuity the consis-
tency of scaling has not been more considered. Whereas the world
vigorously pursues a uniformity for mintage, weight, and measure
of length, it seems to be the other way around in our field, trying
to make the notations as diversified as possible.”11 Just a few years
later, logarithmic progression and the decimal notation, as well as
the Snellen notation, became the accepted standards in clinical
routine worldwide. For reading charts, however, it seems that the
statement of Snellen and Landolt is still true. Jaeger, Nieden,
Parinaud, Decimal, Snellen, M-size, the N notation, line numbers,
logMAR, logRAD, and VAR (Visual Acuity Rating) notations are cur-
rently in use, depending on the users' location and educational
background. It is not within the bounds of this review to present
and discuss all of these notations in detail as has done previously.4
For distance acuity, logMAR has become the worldwide standard
in research.52,56 It is therefore not unlikely that logMAR will once
also predominate in clinical practice. In research on reading acuity,
both logMAR and logRAD (log-reading acuity determination) nota-
tions are both used.4,18 With the RADNER Reading Charts, we intro-
duced the use of logRAD,4,14–18 the reading equivalent of logMAR,
because it was found to be convenient to use different notations
for distance acuity (logMAR) and reading acuity (logRAD). This ap-
proach also facilitates data analysis and the reading of publications.
Another relevant point is that, from a psychophysical perspective,
reading acuity involves a different visual task than does single-
optotype distance acuity,4,14,16–18 further supporting the suggestion
that different definitions be used for the two different tasks. More
widespread use of the logRAD notation would give reading acuity
its own well-deserved research identity.4

ARTICLE INFORMATION REFERENCES Measurement Standard. Ital J Ophthalmol 1988;11:

5–19.

Submitted: March 19, 2019
Accepted: July 13, 2019
Funding/Support: None of the authors have reported
funding/support.
Conflict of Interest Disclosure: WR receives royalties for
the RADNER Reading Charts.
Author Contributions and Acknowledgments: Conceptualization:
WR; Formal Analysis: WR; Methodology: WR; Supervision:
WR; Validation: WR; Writing – Original Draft: WR.
The author would like to thank Prof. Susan Leat, PhD,
Balsam Alabdulkader, BScOptom, MSc, PhD; Lei Liu,
PhD; and Cong Yu, PhD, for their valuable contributions
in writing the paragraphs explaining their reading charts
and for their factual input with regard to the context of
the scientific background. He would also like to thank
Prof. August Colenbrander, MD, PhD, for his valuable
contribution in writing the paragraph about the stages of
visual perception and Deborah McClellan, PhD, for
editorial assistance.
1. Cicchetti D. Guidelines, Criteria, and Rules of Thumb
for Evaluating Normed and Standardized Assessment
Instruments in Psychology. Psychol Assessm 2016;6:
284–90.
2. Fischer R, Milfont T. Standardization in Psychological
Research. Int J Psychol Res 2010;3:81–96.
3. Bortz J, Doering N. Methods for Research and Evalu-
ation for Human and Social Scientists. 4th ed.
Heidelberg: Spinger Verlag; 2006.
4. Radner W. Reading Charts in Ophthalmology. Graefes
Arch Clin Exp Ophthalmol 2017;255:1465–82.
5. Birkhäuser R. Reading Probes for Near Vision from
the University Eye Clinic Bern. Berlin: Julius Springer
Verlag; 1911.
6. Sloan LL, Brown DJ. Reading Cards for Selection of
Optical Aids for Partially Sighted. Am J Ophthalmol
1963;55:1187–99.
7. Law FW. Standardization of Reading Types. Br J
Ophthalmol 1951;35:765–73.
8. Colenbrander A. Consilium Ophthalmologicum
Universale Visual Functions Committee, Visual Acuity
9. Jaeger E. About Cataract and Cataract Surgery Including
Observations and Experiences. Vienna, Austria: Seidel und
Sohn Publisher:1854.
10. Jaeger E. Reading Charts. Vienna, Austria: Verlag
Seidel und Sohn and Victor Masson Publisher; 1856.
11. Graefe A, Saemisch T. Handbook of Ophthalmology.
Leipzig, Germany: Engelmann Publisher; 1874.
12. Nieden A. Reading Probes for the Determination of
Visual Acuity. Wiesbaden, Germany: J. F. Bergman
Publisher; 1882.
13. Radner W. Ophthalmic Reading Tests: Part 1:
Historical Aspects. Ophthalmologe 2016;113:918–24.
14. Radner W, Willinger U, Obermayer W, et al. A New
Reading Chart for Simultaneous Determination of Reading
Vision and Reading Speed. Klin Monbl Augenheilkd
1998;213:174–81.
15. Radner W, Obermayer W, Richter-Mueksch S, et al.
The Validity and Reliability of Short German Sentences
for Measuring Reading Speed. Graefes Arch Clin Exp
Ophthalmol 2002;240:461–7.
16. Stifter E, Konig F, Lang T, et al. Reliability of a Stan-
dardized Reading Chart System: Variance Component

www.optvissci.com Optom Vis Sci 2019; Vol 96(10) 778

Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.


Analysis, Test-retest and Inter-chart Reliability. Graefes
Arch Clin Exp Ophthalmol 2004;242:31–9.
17. Radner W, Diendorfer G. English Sentence
Optotypes for Measuring Reading Acuity and Speed—The
English Version of the Radner Reading Charts. Graefes Arch
Clin Exp Ophthalmol 2014;252:1297–303.
18. Radner W. Near Vision Examination in Presbyopia
Patients: Do We Need Good Homologated Near Vision
Charts? Eye Vis (Lond) 2016;3:29.
19. Maaijwee K, Mulder P, Radner W, et al. Reliability
Testing of the Dutch Version of the Radner Reading
Charts. Optom Vis Sci 2008;85:353–8.
20. Alabdulkader B, Leat SJ. A Standardized Arabic
Reading Acuity Chart: The Balsam Alabdulkader-Leat
Chart. Optom Vis Sci 2017;94:807–16.
21. Alabdulkader B, Leat SJ. Toward Developing a Stan-
dardized Arabic Continuous Text Reading Chart. J Optom
2017;10:84–94.
22. Cheung J, Liu D, Lam C, et al. Development and Val-
idation of a New Chinese Reading Chart for Children.
Ophthalmic Physiol Opt 2015;35:514–21.
23. Han Q, Cong L, Yu C, et al. Developing a Logarithmic
Chinese Reading Acuity Chart. Optom Vis Sci 2017;94:
714–24.
24. Bailey I, Lovie J. The Design and Use of a New
Near-vision Chart. Am J Optom Physiol Opt 1980;57:
378–87.
25. Mansfield J, Ahn SJ, Legge GE, et al. A New
Reading-acuity Chart for Normal and Low Vision. In:
Optical Society of America (OSA) Technical Digest
Series: Ophthalmic and Visual Optics/Noninvasive
Assessment of the Visual System, vol. 3. St. Louis:
OSA; 1993:232–5.
26. MacKeben M, Nair U, Walker L, et al. Random Word
Recognition Chart Helps Scotoma Assessment in Low
Vision. Optom Vis Sci 2015;92:421–8.
27. Radner W, Radner S, Diendorfer G. Integrating a
Novel Concept of Sentence Optotypes into the Radner
Reading Charts. Br J Ophthalmol 2017;101:239–43.
28. Maaijwee KJ, Meulendijks CF, Radner W, et al. The
Dutch Version of the Radner Reading Chart for Assessing Vi-
sion Function. Ned Tijdschr Geneeskd 2007;151:2494–7.
29. Alió JL, Radner W, Plaza-Puche AB, et al. Design of
Short Spanish Sentences for Measuring Reading
www.optvissci.com
Copyright © American Academy of Optometry. Unauthorized reproduction of this article is prohibited.
Performance: Radner-Vissum Test. J Cataract Refract
Surg 2008;34:638–42.
30. Vámosi P, Baló A, Serfó´zó´ C, et al. A New Standard-
ized Reading Chart for the Evaluation of Reading Acuity
and Speed (Hungarian Version of the Radner Reading
Charts). Szemeszet 2009;146:59–63.
31. Ahl J. Measurement of Functional Visual Quality — Radner
Reading Charts in Swedish. Master thesis. Stockholm:
Karolinska Institute; 2007.
32. Calossi A, Boccardo L, Fossetti A, et al. Design of
Short Italian Sentences to Assess Near Vision Performance.
J Optom 2014;7:203–9.
33. Rosa AM, Farinha CL, Radner W, et al. Development
of the Portuguese Version of a Standardized Reading
Test: The Radner-Coimbra Charts. Arq Bras Oftalmol
2016;79:238–42.
34. Munch IC, Jørgensen AH, Radner W. The Danish
Version of the Radner Reading Chart: Design and Empirical
Testing of Sentence Optotypes in Subjects of Varying
Educational Background. Acta Ophthalmol 2016;94:
182–6.
35. Bailey IL, Lovie-Kitchin JE. Visual Acuity Testing.
From the Laboratory to the Clinic. Vision Res 2013;90:
2–9.
36. Legge GE, Ross JA, Luebker A, et al. Psychophysics
of Reading. VIII. The Minnesota Low-vision Reading
Test. Optom Vis Sci 1989;66:843–53.
37. Mansfield JS, Legge GE. The MNREAD Acuity
Chart. In: Legge GE, ed. The Psychophysics of Reading
in Normal and Low Vision. Mahwah, NJ: Lawrence
Erlbaum Associates; 2007:167–91.
38. Carver R. Evidence for the Invalidity of the Miller-Coleman
Readability Scale. J Read Behav 1972;4:42–7.
39. Carver R. Word Length, Prose Difficulty, and Reading
Rate. J Lit Res 1976;8:193–203.
40. Bland JM, Altman DG. Statistical Methods for
Assessing Agreement between Two Methods of Clinical
Measurement. Lancet 1986;1(8476):307–10.
41. Virgili G, Cordaro C, Bigoni A, et al. Reading
Acuity in Children: Evaluation and Reliability Using
MNREAD Charts. Invest Ophthalmol Vis Sci 2004;
45:3349–54.
42. Subramanian A, Pardhan S. The Repeatability of
MNREAD Acuity Charts and Variability at Different Test
Distances. Optom Vis Sci 2006;83:572–6.
Optom Vis Sci 2019; Vol 96(10)
Reading Chart Standardization — Radner
43. Subramanian A, Pardhan S. Repeatability of Reading
Ability Indices in Subjects with Impaired Vision. Invest
Ophthalmol Vis Sci 2009;50:3643–7.
44. Patel PJ, Chen FK, Da Cruz L, et al. Test-retest
Variability of Reading Performance Metrics Using MNREAD
in Patients with Age-related Macular Degeneration. Invest
Ophthalmol Vis Sci 2011;52:3854–9.
45. Trauzettel-Klosinski S, Dietz K, IReST Study Group.
Standardized Assessment of Reading Performance: The
New International Reading Speed Texts IReST. Invest
Ophthalmol Vis Sci 2012;53:5452–61.
46. Lovie-Kitchin JE, Brown B. Repeatability and Inter-
correlations of Standard Vision Tests as a Function of
Age. Optom Vis Sci 2000;77:412–20.
47. Colenbrander A. Vision Rehabilitation Is Part of
AMD Care. Vision 2018;2:4.
48. Flom MC, Weymouth FW, Kahneman D. Visual
Resolution and Contour Interaction. J Opt Soc Am
1963;53:1026–32.
49. Legge GE, Ross JA, Isenberg LM, et al. Psychophys-
ics of Reading. Clinical Predictors of Low-vision Reading
Speed. Invest Ophthalmol Vis Sci 1992;33:677–87.
50. McClure ME, Hart PM, Jackson AJ, et al. Macular
Degeneration: Do Conventional Measurements of Im-
paired Visual Function Equate with Visual Disability?
Br J Ophthalmol 2000;84:244–50.
51. Calabrèse A, Cheong A, Cheung S, et al. Baseline
MNREAD Measures for Normally Sighted Subjects from
Childhood to Old Age. Invest Ophthalmol Vis Sci 2016;
57:3836–43.
52. Ferris FL, 3rd, Kassoff A, Bresnick GH, et al.
New Visual Acuity Charts for Clinical Research. Am J
Ophthalmol 1982;94:91–6.
53. Radner W, Benesch T. Age-related Changes in Base-
line Reading Acuity and Speed as Measured Using
RADNER Reading Charts in Healthy Eyes with Best
Corrected ETDRS Distance Acuity. Br J Ophthalmol
2019;103:1518–23.
54. Whittaker SG, Lovie-Kitchin J. Visual Requirements
for Reading. Optom Vis Sci 1993;70:54–65.
55. Carkeet A. Modeling logMAR Visual Acuity Scores:
Effects of Termination Rules and Alternative Forced-
choice Options. Optom Vis Sci 2001;78:529–38.
56. Bailey IL, Lovie JE. New Design Principles for Visual Acu-
ity Letter Charts. Am J Optom Physiol Opt 1976;53:740–5.
779