|Year : 2018 | Volume
| Issue : 5 | Page : 630-633
Reliability and validity of a Kannada rate of reading test
Krithica Srinivasan1, Gopee Krishnan2, Arnold Wilkins3, Peter Allen4
1 Department of Optometry, School of Allied Health Sciences, Manipal Academy of Higher Education, Manipal, Karnataka, India
2 Department of Speech and Hearing, School of Allied Health Sciences, Manipal Academy of Higher Education, Manipal, Karnataka, India
3 Department of Psychology, University of Essex, Colchester, UK
4 Department of Vision and Hearing Sciences, Vision and Eye Research Unit, Anglia Ruskin University, Cambridge, UK
|Date of Submission||19-Aug-2017|
|Date of Acceptance||02-Feb-2018|
|Date of Web Publication||20-Apr-2018|
Dr. Peter Allen
Department of Vision and Hearing Sciences, Vision and Eye Research Unit, Anglia Ruskin University, East Road, Cambridge CB1 1PT
Source of Support: None, Conflict of Interest: None
Purpose: Kannada, one of the Dravidian languages, is the official language of Karnataka state of India. There is a need for a test using Kannada words that can assess visual aspects of reading independently of syntactic and semantic knowledge. Methods: A test of reading rate in Kannada was developed following the design principles of the Wilkins Rate of Reading Test (RRT). Fifteen high-frequency bisyllabic Kannada words were selected. Children were recruited from state and private schools that used Kannada or English as the medium of instruction. A total of 799 children from Grade 2 to 9 participated in the study. Reading rate was measured using the English RRT and the Kannada version twice in immediate succession during the first session. In 85 children, measurements using the Kannada RRT were repeated after an interval of 15 days. Results: Pearson product moment correlation between the two immediately successive tests was 0.95 for the Kannada RRT and 0.91 for the English RRT. The correlation for the tests separated by an interval of 15 days was 0.83. When Kannada was the medium of instruction, there was little difference between test scores for Kannada and English. When English was the medium of instruction, test scores were greater in English. Scores increased as expected with age (P < 0.0001), similarly for Kannada and English tests. Conclusion: The newly developed Kannada RRT is both reliable and valid and can be used as a tool for measuring the visual aspects of reading.
Keywords: English language, Kannada language, rate of reading test
|How to cite this article:|
Srinivasan K, Krishnan G, Wilkins A, Allen P. Reliability and validity of a Kannada rate of reading test. Indian J Ophthalmol 2018;66:630-3
|How to cite this URL:|
Srinivasan K, Krishnan G, Wilkins A, Allen P. Reliability and validity of a Kannada rate of reading test. Indian J Ophthalmol [serial online] 2018 [cited 2018 May 24];66:630-3. Available from: http://www.ijo.in/text.asp?2018/66/5/630/230683
Reading is a complex task involving vision, phonological decoding, word recognition, and comprehension. In schools, it is typically assessed by asking pupils to read short passages either aloud or silently, perhaps answering questions on the content of the passage. If a pupil underperforms, it is not clear without further testing which of the various component skills, such as vision, is the source of reading difficulty.
Traditionally, vision is measured in an eye care clinic by requiring an individual to read small isolated letters. In conventional text, the letters are larger than the acuity limit but in close proximity to one another, crowded into words. Further, the binocular control required for reading text is more complex than that required for the letter chart, involving the reestablishment of appropriate vergence after each saccade. The letter strokes in the text can be similar to one another, impairing this vergence control. The lines of text both from the vertical letter strokes and from the lines of words resemble patterns of stripes that can cause perceptual distortions and discomfort, now collectively referred to as visual stress.,
The visual stress is sometimes treated using colored sheets of plastic placed on text when reading (colored overlays). The Wilkins Rate of Reading Test (RRT) was developed to measure the effect of overlays in improving reading speed. It assesses the visual aspects of reading independently of the syntactic and semantic knowledge, which would otherwise increase the variance in reading speed. The test consists of a paragraph of randomly ordered common words that are read aloud as quickly as possible, and it is scored in terms of the number of words correctly read in 1 min. The test has proved useful in identifying the individuals who will subsequently continue to use overlays voluntarily in the long term. The test can also assess the effects on reading speed of refractive errors and prisms for decompensated heterophoria., The purpose of the present study was to create a version of the test in the Kannada language, used by people in the southwest of India, which has a transparent orthography, unlike English. Performance on the Kannada version was compared with performance on the English version in children across different grades with Kannada or English as the medium of instruction. Repeated measures of both the English and Kannada versions were performed to estimate the reliability.
| Methods|| |
This study involved the development of a Kannada RRT to assess visual aspects of reading and facilitate estimation of its reliability and validity. All participants provided informed consent, and the research was conducted in accordance with the tenets of the Declaration of Helsinki. The study was approved by the Institutional Research and Ethics Committees.
Development of Kannada reading rate test
A group of 97 Kannada words with a word frequency (frequency of occurrence) ≥ 90/100,000 words were selected from a repository of words published by the Central Institute of Indian Languages. Only 3 of the 97 words were monosyllabic, 65 were bisyllabic, and the remaining words were either trisyllabic or tetrasyllabic. Owing to the paucity of monosyllabic words, 15 bisyllabic “content and nonpicture” words were selected for the development of the Kannada Reading Rate Test (Kannada RRT), including verbs, nouns, adjectives, or adverbs. The selected words are given in [Table 1].
The Kannada RRT was based on the same design principles as the English RRT. Four versions of the Kannada RRT were developed. Each version had 10 lines consisting of the 15 selected Kannada words in a different random order. Each line was checked to ensure that (1) the words did not combine to form meaning; (2) same word was not used at the end of one line and the beginning of the next, and (3) no consecutive lines had the same words vertically aligned.
“Tunga” font of 8-point size was used, and the reading paragraph was designed using Microsoft Publisher. This size was selected to match the vertical x-height of the English version. All words were single-spaced horizontally. The spacing between the lines was 0.8 point and between words was 0.65 point. The four different versions were printed as black letters on a white background using a 600 dpi laser printer. [Figure 1] shows one of the versions of Kannada RRT.
Participants were children from Grades 2 to 9 recruited from 9 schools located in Udupi Taluk including 4 private and 5 state schools that had English and Kannada, respectively, as the principal medium of instruction. Both males and females from each grade who were present on the day of examination were considered for the study. Based on the inclusion and exclusion criteria, 430 boys and 369 girls were recruited, ranging in age from 7 to 16 years. [Figure 2] shows the inclusion and exclusion criterion alongside participant numbers at each stage of the study. Only three schools provided permission for conducting a repeat test. Based on simple random sampling, 85 children from these schools who were present on the day of testing (15 days after the initial assessment) were included for the repeat test. They included 57 boys and 28 girls, mean (standard deviation [SD]) age of 12 (1.6) years, and range: 9–16 years. The testing rooms were illuminated by daylight to a minimum of 600 lux with a maximum of 900 lux.
|Figure 2: Inclusion and exclusion criterion alongside participant numbers at each stage of the study|
Click here to view
Rate of reading
Participants completed the Kannada and English versions of the RRT twice during the first session as part of a study of the effects of a colored overlay to be reported in due course. Each passage was read aloud for 1 min (timed with a stopwatch by KS), and the number of words read correctly was recorded in words per minute (wpm). The Administration of English and Kannada versions were alternated across participants. The RRT was administered for all participants at a viewing distance between 0.4 and 0.5 m. During the second session, only the Kannada RRT was repeated.
Reliability was assessed using Pearson product moment correlations between the first and second scores on the RRTs. Linear regression analysis was used to analyze the relationship between age and rate of reading.
| Results|| |
The test was administered twice in immediate succession during the first session in English and Kannada. The Pearson product moment correlation between the two test results performed during the first session was 0.95 (P< 0.001) for the Kannada RRT and 0.91 (P< 0.001) for the English RRT. For 85 children, the tests were administered a second time after 15 days. The correlation between the first test on the two sessions was 0.83 (P< 0.001).
The relation between the two test scores in the first session is shown separately for the Kannada and English tests in [Figure 3]a and [Figure 3]b.
|Figure 3: A scatter plot showing the scores in words per minute on the second test as a function of scores on the first test on the (a) Kannada rate of reading test and (b) the English rate of reading test. The R2 values in the figures show the proportion of variance explained by the relationship between the two scores|
Click here to view
The Kannada test showed a small practice effect: the second test was read 1% more quickly than the first (t(798) = 5.32, P < 0.001). In the English version, there was more variation and a weak nonsignificant tendency for the second score to be lower than the first. For both tests, there was a weak but significant tendency for the difference between first and second tests to be smaller for children with high scores (t(798) = 4.57, P < 0.001, for the Kannada test and t(798) = 6.66, P < 0.001 for the English version).
The absolute value of the difference between the two scores was expressed as a proportion of their mean. For the Kannada version of the test, the proportional difference averaged 7.7% (SD 7.1%). For the English version, the values were 7.2% (SD 8.3%). Differences of 15% are therefore more than 2 SDs from the mean.
[Table 2] shows the mean and SD for the rate of reading for each age group and each test.
|Table 2: Mean and standard deviation for the rate of reading (wpm) for each age group and each test|
Click here to view
The test scores increased with age as shown in [Figure 4].
|Figure 4: Rate of reading as a function of age on the Kannada rate of reading test (solid lines) and English rate of reading test (broken lines) when the medium of instruction was Kannada (solid symbols) and English (open symbols)|
Click here to view
The linear regressions on age were all very highly significant (P< 0.0001), accounting for a minimum of 61% of the variance. There was no overall effect of gender on reading rate, and so in [Figure 3] and [Figure 4], the data for boys and girls have been combined.
| Discussion|| |
In the southwest zone of India, Kannada is the official language of Karnataka state, with more than 40% of school children having their medium of education in Kannada. The Kannada RRT was developed for use in this zone. The Kannada RRT had a reliability similar to that previously reported for the English version.
The intraparticipant variation for the Kannada RRT was similar to that for the English version indicating that an increase of >15% in rate of reading is greater than might be expected from chance variation and therefore clinically significant.,
The progression in reading rate with age indicated that the RRT was a valid measure of reading skill and sensitive to reading development. The rate of development was similar for the English and Kannada versions of the test despite differences in the language of instruction.
In India, most of the state schools have their regional language as the medium of instruction whereas private schools use English. Students from private schools have been shown to have generally better language skills., The proficiency of teachers in private schools has been estimated to be better than those in state schools. The overall differences in reading rate among children in state and private schools are consistent with these differences and provide a further indication of the validity of the test.
The Kannada RRT is designed to measure the effects on reading of visuoperceptual distortions of text, such as apparent movement or blurring of words and letters. It is therefore proposed to use the test as an adjunct in primary eyecare clinics among children complaining of visuoperceptual difficulties, to assess the immediate effect of any intervention be that refractive, vision therapy, and/or colored overlays. It is worth noting that the Kannada RRT is not a conventional reading test in which the words become progressively more difficult and so does not measure “reading age;” but instead, it assesses reading ability in terms of rate and errors rather than in terms of the difficulty of the words read. Moreover, there is no evidence that individuals who read faster do so because they tolerate a greater number of errors. In general, the conditions that give rise to errors also give rise to slower reading.
| Conclusion|| |
There are numerous different reading tests available that assess the educational aspects of reading performance. The RRT is designed to measure the visual aspects of reading and can assess any visual deficits that might affect reading. We have provided a version for use with the Kannada language.
Financial support and sponsorship
Peter Allen's participation was supported by a Collaborative Research Grant from the College of Optometrists.
Conflicts of interest
There are no conflicts of interest.
| References|| |
Jainta S, Jaschinski W, Wilkins AJ. Periodic letter strokes within a word affect fixation disparity during reading. J Vis 2010;10:2.
Wilkins A, Milroy R, Nimmo-Smith I, Wright A, Tyrrell R, Holland K, et al.
Preliminary observations concerning treatment of visual discomfort and associated perceptual distortion. Ophthalmic Physiol Opt 1992;12:257-63.
Wilkins AJ, Nimmo-Smith MI. The clarity and comfort of printed text. Ergonomics 1987;30:1705-20.
Wilkins AJ, Jeanes RJ, Pumfrey PD, Laskier M. Rate of reading test: Its reliability, and its validity in the assessment of the effects of coloured overlays. Ophthalmic Physiol Opt 1996;16:491-7.
O'Leary CI, Evans BJ, Edgar DF. The effect of low refractive corrections on rate of reading. Optom Pract 2014;15:87-100.
O'Leary CI, Evans BJ. Double-masked randomised placebo-controlled trial of the effect of prismatic corrections on rate of reading and the relationship with symptoms. Ophthalmic Physiol Opt 2006;26:555-65.
Ranganatha MR. Morphophonemic Analysis of the Kannada Language: Relative Frequency of Phonemes and Morphemes in Kannada. Central Institute of Indian Languages Occasional Monographs Series 17. Mysore: Central Institute of Indian Languages; 1982.
Wilkins AJ, Allen PM, Monger L, Gilchrist JM. Visual stress and dyslexia for the practising optometrist. Optom Pract 2016;17:103-12.
Anand CL. The Study of the Effect of Socio-Economic Environment and Medium of Instruction on the Mental Abilities and Academic Achievement of Children in Mysore State. Ph.D. Dissertation, University of Mysore; 1971.
Srivastava AL, Khatoon R. Effect of Difference Between Mother Tongue and Another Tongue as Medium of Instruction on Achievement, Mental Ability and Creativity of the Vll Standard Children. Bilingualism and Achievement in Schools. Mysore: CIIL; 1980.
Shende SN. A comparative study of English language teaching in English and Non-English medium secondary schools/colleges from Grade 8 to Grade 12 in India. Lang India 2015;15:157-75.
[Figure 1], [Figure 2], [Figure 3], [Figure 4]
[Table 1], [Table 2]