The optometrist often encounters school-aged children with visual complaints related to reading and near work.1 For example, children with convergence insufficiency have symptoms associated with reading, such as loss of place and difficulty comprehending text.2,3 One difficulty for the optometrist is separating visual factors that adversely impact reading from language-based factors. Research has found that poor decoding skills are often associated with reading difficulties and have their origin in specific language-based skills.4–7 To help address this issue, Griffin and Walton8 developed the Dyslexia Determination Test (DDT) as a method to diagnose difficulties in single word reading related to phonetic word attack skills and sight-word recognition. Children who struggle with phonetic word analysis have difficulty using grapheme-to-phoneme relationships, which leads to difficulty reading unfamiliar words. Those who have difficulty with sight-word recognition have difficulty perceiving words as visual gestalts but can decode phonetically regular words and tend to be slow readers. A combination of the two types of decoding problems is very disabling for children learning to read. The DDT is a comprehensive test of assessing single word reading that takes 15 to 20 minutes to administer. Griffin et al. also developed a shortened version of the test to be used as a screening tool called the Decoding-Encoding Screener for Dyslexia (DESD). The DESD was developed by altering and expanding an older screening test: the Dyslexia Screener. Unlike the Dyslexia Screener, which took words directly from the DDT, the DESD uses different words for decoding compared with the DDT. It also added an analysis of word difficulty to rank words by grade level, expanded the reliability and validity-related evidence, and added an additional word reading test. As a result, the DESD may be more appropriate than the DDT for an optometrist who needs information regarding single word reading level as well as decoding skills.9
As compared with previous dyslexia tests developed by Griffin, the DESD is unique in that it uses the results of the Decoding section, referred to as the Reading Test, to approximate the reading level of the child. The Reading Test standard score result allows comparison of a child’s performance to standard scores from other tests of reading. In addition to the Reading Test, the DESD is composed of an optional Letter Writing section (not included in this study) and an Encoding section that uses a Sight-Word Spelling Test and a Phonetic Spelling Test to rank the child’s encoding ability. The results of the DESD determine those subjects who may have reading difficulties and may need more comprehensive testing.
Despite the improvements in the DESD, there is limited validity-related evidence for the Reading Test and the Sight-Word and Phonetic Spelling Tests. The DESD technical manual summarizes the current measures of validity for the DESD. The DESD Reading Test raw score and the Sight-Word and Phonetic Spelling Tests were correlated with the word recognition test of the Norris Educational Achievement Test (NEAT). The correlation with the DESD Reading Test raw score was high (0.87), but a lower correlation was found for the Sight-Word Spelling Test (0.54). A second study showed that the DESD Reading Test raw score correlation with the NEAT was also high (0.85), but lower correlations were found for the Sight-Word (0.48) and the Phonetic Spelling Tests (0.65). A study comparing the DESD Reading Test raw score with the Iowa Test of Basic Skills (ITBS) found high correlations for the Vocabulary (0.75) and the Comprehension (0.72) Tests. Thus, there is limited validity-related evidence for the DESD test.9
The aim of this study was to evaluate the validity of the DESD Reading Test and the Sight-Word and Phonetic Spelling Tests using two widely accepted reading assessment tests: the Wide Range Achievement Test-Third Edition (WRAT-3) and the Gray Oral Reading Test-Fourth Edition (GORT-4).
This study was approved by the institutional review board at the Southern California College of Optometry for the protection of human subjects. Written consent was obtained from the parents/guardians, and child written assent was obtained for all study participants.
Children aged 9 to 15 years were recruited from the University Eye Center of Fullerton, CA, and from parents who were faculty, staff, or professional students at the Southern California College of Optometry. Major eligibility criteria included the following: English as a primary language, visual acuity of 20/25 or better at distance and near, lateral disparity stereopsis of 70 seconds of arc or better, and no strabismus or amblyopia. Children with uncorrected refractive error (determined by noncycloplegic refraction) of the following amounts were excluded: greater than 0.50 diopters (D) myopia, greater than 1.00 D astigmatism, greater than 1.00 D anisometropia, and greater than 1.50 D hyperopia.
Three reading tests, the DESD, the WRAT-3, and the GORT-4, were administered according to procedures outlined in the respective manuals. The order of administration was randomized, and children were given a 1- to 2-minute break between each test.
The DESD consists of Decoding (Reading Test) and Encoding Sections (Sight-Word and Phonetic Spelling Tests).9 The performance on the Reading Test determined the words used for the Sight-Word and Phonetic Spelling Tests. The child was instructed to read aloud up to 55 words arranged in order of increasing difficulty and divided into groups of five words corresponding to grade levels from kindergarten to college level. Children in grades 1 to 4 began reading words at the kindergarten level, and older children began reading words three grade levels below their grade placement. The Reading Test raw score was based on the number of words read correctly and converted to a standard score based on the child’s grade level. For the Sight-Word Spelling Test, five words were dictated starting at the DESD grade level, the highest level at which the subject was able to correctly read three of the five words. The words selected started with the highest numbered word the child read correctly at the DESD grade level and worked backward toward lower numbered words pronounced correctly. The words were read aloud by the examiner one at a time and the child was asked to spell the words correctly on the form. For the Phonetic Spelling Test, five words were dictated starting one grade level above the child’s DESD grade level with the lowest numbered word the child missed and worked toward higher numbered words. Each word was read aloud by the examiner one at a time and the child was asked to write the word the way it sounds. The Sight-Word and Phonetic Spelling Test raw scores were based on percentage correct out of five words. The classification of performance from markedly below normal to above normal was determined by the child’s actual grade placement relative to the DESD grade level and the spelling test raw scores.
The WRAT-3 is an assessment tool that measures the basic skills of reading, spelling, and arithmetic (not used in this study) for subjects aged 5 to 75 years.10 It assesses single word reading and spelling ability and would be expected to parallel the DESD Reading Test and the Sight-Word and Phonetic Spelling Tests. The Word Reading subtest was administered first, followed by the Spelling subtest. The Word Reading subtest required the child to decode a series of words of increasing difficulty until 10 consecutive words were mispronounced or unknown. In the Spelling subtest, words were read out loud to the child according to the pronunciation guide in the testing manual. The child was given 15 seconds to spell each word, and the test was discontinued after 10 consecutive words were spelled incorrectly. For the Reading and Spelling subtests, a standard score was derived based on the chronological age of the child.
In contrast, the GORT-4 assesses a broader range of reading skills including comprehension.11 A comparison of the GORT-4 with the DESD would determine whether a test of single word decoding (DESD) would be a good predictor of performance when reading typical text for comprehension. The GORT-4 assesses oral reading ability in subjects aged 6 to 19 years based on reading passages of increasing difficulty.
The child read each passage aloud and then answered five multiple choice questions about the content of each story. During oral reading, errors of spoken versus written words (Accuracy) and the time in seconds to complete the story (Rate) were recorded. The test provides Rate, Accuracy, Fluency (combination of the rate and accuracy score), and Comprehension scaled scores (mean of 10 and SD of 3). The Fluency and Comprehension scores are combined to form an Oral Reading Quotient standard score (mean of 100 and SD of 15).
This study investigated the associations between the DESD Reading Test and DESD Spelling Tests with the WRAT-3 and the GORT-4. We compared the standard scores for the DESD Reading Test with the standard scores for the WRAT-3 and scaled scores for the GORT-4 Tests using Pearson correlation coefficient.
The DESD Sight-Word and Phonetic Spelling Tests rank a child’s ability from markedly below normal to above normal and do not provide standard scores. To compare these values to standard scores, we assigned a numerical value for the rankings as follows: markedly below normal (1), moderately below normal (2), mildly below normal (3), borderline (4), normal (5), and above normal (6). The scores from the Sight-Word and Phonetic Spelling Tests were combined into a total numerical value referred to as the “DESD Spelling Scale” in this article. The DESD Spelling Scale could range from 2 to 12. To compare the ordinal ranking of the DESD Spelling Scale with the standard scores from the WRAT-3 and GORT-4, we used Spearman rank correlation coefficient.
To assess the clinical significance of correlations, we used a system developed by Hopkins12 that classifies the correlations as follows: 0.0 to 0.1, very small; 0.1 to 0.3, small; 0.3 to 0.5, moderate; 0.5 to 0.7, large; 0.7 to 0.9, very large; and 0.9 to 1.0, nearly perfect.
To assess for possibilities of systematic bias between reading tests, we evaluated the distribution of the differences between test measurements using the method developed by Bland and Altman that calculates mean differences and the 95% limits of agreement. That is, the difference between measurements 1 and 2 is plotted versus the mean of measurements 1 and 2. We conducted this analysis for the DESD Reading Test and the GORT-4 Oral Reading Quotient and the WRAT-3 Word Reading and Spelling subtests. We could not make comparisons to the Rate, Accuracy, Fluency, and Comprehension scores from the GORT-4 owing to the use of scaled scores (mean of 10 and SD of 3) for these subtests.
Thirty-one children aged 9 to 15 years participated in the study with a mean (±SD) age of 11.5 (±2.0) years. Twenty-one were male and 10 were female. The mean (±SD) standard score for the DESD Reading Test was 94.33 (±18.30) (range, 47 to 126). The mean (±SD) score for DESD Spelling Scale was 8.40 (SD = 3.12, range 3 to 12). The WRAT-3 Word Reading subtest standard score mean (±SD) was 98.90 (±10.97) (range, 72 to 127), and the WRAT-3 Spelling subtest standard score mean (±SD) was 93.83 (±15.33) (range, 64 to 120). The mean (±SD) scaled scores for the GORT-4 subtests of Rate, Accuracy, Fluency, and Comprehension were 8.35 (±3.50) (range, 2 to 15), 8.19 (±2.94) (range, 2 to 13), 7.42 (±3.64) (range, 1 to 13), and 9.68 (±2.71) (range, 4 to 15), respectively. The mean (±SD) Oral Reading Quotient score for the GORT-4 was 92.35 (±18.06) (range, 55 to 116).
The Pearson correlation coefficients between the DESD Reading Test and WRAT-3 Spelling subtest and all subtests of the GORT-4 were significant (p < 0.001) and were classified as large. However, the DESD Reading Test had a small, nonsignificant (p = 0.199) correlation with the WRAT-3 Word Reading subtest. The correlations are displayed in Fig. 1A to G.
Using Spearman rank correlation coefficients, the DESD Spelling Scale had statistically significant correlations with the Word Reading and Spelling subtests from the WRAT-3 and all scores from the GORT-4 (p < 0.001). The correlations were large to very large with the WRAT-3 Spelling subtest and all of the GORT-4 subtests and the Oral Reading Quotient. However, the correlation with the WRAT-3 Word Reading subtest was lower with a moderate correlation. The correlations are displayed in Fig. 2A to G.
The Bland-Altman analysis comparing the DESD Reading Test and the WRAT-3 Word Reading and Spelling subtests showed minimal bias. The DESD Reading Test score was 3.50 standard score points lower than the WRAT-3 Word Reading subtest score and 1.52 standard score points above the WRAT-3 Spelling subtest score. A similar result was observed for the GORT-4 Oral Reading Quotient where the DESD Reading Test was 3 standard score points higher (Fig. 3A to C).
The results of the study suggest that the DESD Reading Test and the DESD Spelling Scale showed large associations with the WRAT-3 Spelling subtest and all measures from the GORT-4. However, the association between the DESD Reading Test and the WRAT-3 Word Reading subtest showed a small nonsignificant correlation, and the association with the DESD Spelling Scale showed a moderate association with the WRAT-3 Word Reading subtest.
We compared our results to those reported in the DESD manual and found consistent findings when looking at the ITBS test. The correlations between the Vocabulary and Comprehension sections of the ITBS and the DESD Reading test raw scores were 0.75 and 0.72, but the Spelling Scale correlations were not reported. This is similar to the DESD Reading Test correlations with the GORT-4 that ranged from 0.627 to 0.691 and the DESD Spelling Scale and the GORT-4 that ranged from 0.667 to 0.781. The large correlation between both the DESD Reading Test standard score and the DESD Spelling Scale with the GORT-4 indicates that the DESD, a brief test of decoding and encoding, had a significant relationship with a broad range of oral reading skills, which included fluency and comprehension. This finding is consistent with current models of reading where single word decoding is a major input to measures of reading comprehension.13,14 However, future studies should evaluate alternative measures of comprehension due to criticisms of the GORT-3 comprehension measures.15
We also compared our results to those reported in the DESD manual and found consistent findings when looking at the correlation between the NEAT Word Recognition Test and the DESD Spelling scores. The NEAT correlations with the Spelling scores were 0.48 and 0.65, which are comparable to our correlation of the DESD Spelling Scale with the WRAT-3 Word Reading subtest (0.520) and Spelling subtest (0.666). However, the method used for quantifying scores for the DESD Spelling Tests to compare with the NEAT was not clearly defined in the DESD manual, and we are not sure if our DESD Spelling Scale matches those used by the test’s authors. Thus, our results are similar to the NEAT and ITBS studies reported in the manual, suggesting that the DESD Spelling Scale may be a useful tool for identifying single word reading problems that interfere with reading ability.
In contrast to the above findings, the comparisons of our results to those reported in the DESD manual for the correlation between the DESD Reading Test and the NEAT test were not consistent. The reported correlations in the DESD manual from two different studies comparing the DESD Reading Test and NEAT Word Recognition Score were 0.85 and 0.87. Although the correlation between the DESD Reading Test and the WRAT-3 Spelling subtest was large (0.596), the correlation with the WRAT-3 Word Reading subtest was small (0.237) and well below the values reported for the NEAT. The latter result was unexpected given that both tests evaluate single word decoding and have shown significant correlations with other single word reading tests.9,16,17 In regard to the DESD Reading Test, the manual reports that raw scores were used for the correlations and we used the standard scoring system created for the DESD Reading Test. We reanalyzed our results using raw scores for the DESD Reading Test, and our correlation values were slightly lower than using standard scores from the DESD. A possible explanation for the poor correlation is that the DESD Reading test emphasizes sight-word recognition of phonetically irregular words within 2 seconds whereas the WRAT-3 Word Reading subtest allows the child 10 seconds to decode a mix of phonetically regular and irregular words. As a result, the DESD places more emphasis on sight-word recognition than the WRAT-3 Word Reading subtest, which requires both phonetic and sight-word recognition during testing. At this time, we cannot conclude that the poor correlation between the DESD Reading Test and WRAT-3 Word Reading subtest is attributed to the difference in word type and time given to decode the words. Because of the inconsistent correlations of standardized single word tests with the DESD Reading Test reported in the manual and in our current study, the validity of the DESD Reading Test in identifying single word reading problems is unclear. Future research should compare the DESD Reading Test to other tests of single word decoding to address this issue.
Our study included more male than female subjects; however, none of the tests included in this study provide normative data indicating differences in performance by sex. To further explore sex differences, a two-sample t test comparing sex found no significant differences (p > 0.244) for the three reading tests. We also investigated possible bias between the DESD Reading Test and the WRAT-3 Word Reading and Spelling subtests, and the Oral Reading Quotient and found minimal bias. Thus, bias between tests would not explain the poor correlation between the DESD Reading Test and the WRAT-3 Word Reading subtest. Another possible limitation of our study is the relatively small sample size. However, the sample did include children with a wide range of reading skills.
We did not include emerging readers (aged 6 to 8 years) in our study because of different assessment approaches used in this age group. In early reading, an emphasis is placed on cognitive processes that are foundational for developing adequate decoding skills, and assessments usually include tests of phonological awareness and rapid naming.5–7 In addition, emerging readers who are at risk for developing a reading disability typically have very limited sight-word decoding skills and may not provide a sufficient number of words for the spelling assessment on the DESD. We decided to include children who would be expected to have a sufficient sight-word vocabulary to allow testing with all three tests. We hope to investigate the association between phonological awareness and rapid naming with results from the DESD in emerging readers in future studies.
In conclusion, both the DESD Reading Test and DESD Spelling Scale showed significant correlations for children aged 9 to 15 years with the WRAT-3 Spelling subtest and the GORT-4. However, the small correlation between the DESD Reading Test and the WRAT-3 Word Reading subtest indicates that some caution should be used when interpreting the results of the DESD Reading Test. At this time, the results of the DESD Spelling Scale have better correlations with the WRAT-3 and GORT-4 than the DESD Reading Test and may be a more useful tool for the practitioner.
Midwestern University Eye Institute
19379 N 59th Ave
Glendale, AZ 85308
e-mail: [email protected]
The authors thank Dr. Andrew Loc Nguyen, Loan Trinh, Steve Gildersleeve, and Heidi Peterson for their contributions to this study.
Received October 4, 2014; accepted January 16, 2015.
1. Borsting E, Rouse MW. Detecting learning-related visual problems in the primary care setting. J Am Optom Assoc 1994; 65: 642–50.
2. Borsting EJ, Rouse MW, Mitchell GL, Scheiman M, Cotter SA, Cooper J, Kulp MT, London R. Validity and reliability of the revised convergence insufficiency symptom survey in children aged 9 to 18 years. Optom Vis Sci 2003; 80: 832–8.
3. Borsting E, Rouse MW, Deland PN, Hovett S, Kimura D, Park M, Stephens B. Association of symptoms and convergence and accommodative insufficiency in school-age children. Optometry 2003; 74: 25–34.
4. Stanovich KE. Explaining the differences between the dyslexic and the garden-variety poor reader: the phonological-core variable-difference model. J Learn Disabil 1988; 21: 590–604.
5. Snow CE, Burns MS, Griffin P, eds. Preventing Reading Difficulties in Young Children. Washington, DC: National Academy Press; 1998.
6. Shaywitz SE, Shaywitz BA, Fletcher JM, Escobar MD. Prevalence of reading disability in boys and girls. Results of the Connecticut Longitudinal Study. JAMA 1990; 264: 998–1002.
7. Torgesen JK, Wagner RK. Alternative diagnostic approaches for specific developmental reading disabilities. Learn Disabil Res Prac 1998; 13: 220–32.
8. Griffin JR, Walton HN. Dyslexia Determination Test (DDT), 3rd ed. Optometric Extension Program Foundation: Santa Ana, CA; 2003.
9. Griffin JR, Walton HN, Christenson GN. Decoding-Encoding Screener for Dyslexia (DESD). Western Psychological Services: Los Angeles, CA; 2006.
10. Wilkinson G. Manual for the Wide Range Achievement Test, 3rd ed. Wilmington, DE: Wide Range Inc.; 1993.
11. Wiederholt JL, Bryant BR. Gray Oral Reading Tests, 4th ed. Pro-Ed: Austin, TX; 2001.
12. Hopkins W. A New View of Statistics. A Scale of Magnitudes for the Effect Statistics. Available at: http://www.sportsci.org/resource/stats/effectmag.html
. Accessed August 15, 2014.
13. Catts HW, Adlof SM, Ellis Weismer S. Language deficits in poor comprehenders: a case for the simple view of reading. J Speech Lang Hear Res 2006; 49: 278–93.
14. Tunmer WE, Chapman JW. The simple view of reading redux: vocabulary knowledge and the independent components hypothesis. J Learn Disabil 2012; 45: 453–66.
15. Keenan JM, Meenan CE. Test differences in diagnosing reading comprehension deficits. J Learn Disabil 2014; 47: 125–35.
16. Smith TD, Smith BL. Relationship between the Wide Range Achievement Test 3 and the Wechsler Individual Achievement Test. Psychol Rep 1998; 83: 963–7.
17. Weschler D. Weschler Individual Achievement Test, 2nd ed: (WIAT-IIA). San Antonio, TX: Harcourt Assessment, Inc.; 2005.