258 IER : SPECIAL NUMBER 1995
side for all the three sub-groups and in Kerala the difficulty level appears high only for the urban sample.
Looking at content analysis of language textbooks in terms of Vocabulary (Table 4) and Syntax (Table 5) it was observed that language textbooks of the states show a common pattern with maximum usage of nouns followed by adjectives, verbs and adverbs, the only deviation being in the Maharashtra textbook where percentage of adjectives overtakes percentage of verbs ( Table 4). The above pattern was also reported by Sharma (1993) in his study on comprehensibility of primary level textbooks of Rajasthan. Pattern of children's own vocabulary at Grade III level (Sharma, 1993) was also in tune and thus a positive indication in favour of the language textbooks
The trend of sentence usage was found consistent for the language textbooks of' Haryana, Maharashtra and Karnataka with maximum usage of simple sentences followed by complex and then compound sentences; in Assam, there is a slight deviation. The percentage of compound sentences (6.48) is marginally higher than that of complex sentences (5.96). Interestingly, there were no, compound sentences used in Tamil Nadu and Kerala language readers. But the
259
260 IER : SPECIAL NUMBER 1995
TABLE 4
State and Subject-wise Linguistic Analysis of (Grade III) Textbooks
in Percentages
subject State Noun Adjective Adverb Verb
Language Assam 58.31 5.54 5.79 30.60
Haryana 50.95 16.031 4.44 18.58
Maharashtra 45.45 25.58 5.07 23.90
Karnataka 53.53 7.05 4.36 35.06
Kerala 50.91 16.86 7.77 24.46
Tamil Nadu 54.30 4.30 8.87 32.53
EVS I Assam 66.98 7.15 1.11 24.75
Haryana 55.65 20.93 10.30 13.12
Maharashtra 60.74 197.5 1.98 17.53
Karnataka 82.31 5.38 1.54 10.77
Kerala 62.86 8.57 4.08 24.49
Tamil Nadu 50.52 7.55 9.90 32.03
EVS II Assam 59.58 8.41 1.41 30.60
Haryana 31.45 13.04 16.52 38.99
Maharashtra 57.36 27.93 2.01 12.61
Karnataka 64.33 6.42 4.09 25.15
Kerala 67.89 10.75 1.33 20.03
Tamil Nadu 55.37 10.24 5.85 18.54
Maths Assam 41.76 25.95 10.13 22.16
Haryana 62.74 21.97 2.23 13.03
Maharashtra 54.32 15.14 3.78 26.76
Karnataka 71.58 2.19 1.09 25.14
Kerala 61.12 15.23 5.08 18.27
Tamil Nadu 51.16 7.44 9.77 31.63
percentage of simple sentences remain higher in Kerala whereas in Tamil Nadu percentage of complex sentences overtook simple sentences. Apparently, this factor did not affect the readability of the textbooks as judged by the readability scores of children on language test across different states. Sharma (1993) reported it slightly different pattern in Rajasthan language (Hindi) textbook with highest percentage of simple sentences (88.7) followed by compound (7.23) then complex sentence (4.06) Pattern of sentence usage as observed in spontaneous language of children of Grade III in Sharma's study is more consistent with that of Maharashtra, Haryana and Karnataka with highest percentage of simple sentences followed by complex (7.23) and then compound sentences.
State-wise variations were observed in sentence length/sentence type. In the case of simple sentences average sentences length ranges from 4.10 in Tamil Nadu textbook to 7.70 in Haryana textbook. For complex sentences, length ranges
261
262 IER : SPECIAL NUMBER 1995
from 6.90 in the Tamil textbook to 15.00 in Haryana textbook. In compound sentences length ranges from 9 words in Karnataka textbook to 14.00 in Haryana. To probe further into the factor of sentence type and sentence length, vis-a-vis children's performance a comparison was made of the two state's textbooks found to be highest (Maharashtra) and lowest (Haryana) on readability as per children's means scores (Table 6). It seems that the two factors which may have some bearing on lowering the readability level could be the higher usage of compound sentences (8.47) and higher average sentence length (7.7, 14 and 15 for simple, compound and complex sentences respectively) in Haryana textbook. This becomes particularly relevant against the backdrop of the pattern evident in Grade III children's spontaneous language (Sharma, 1993) where the percentage of compound sentences used is only 4.06 and average sentence length for simple sentence (spoken and written) is 4.65. Maharashtra sentence pattern is closer to this and is therefore possibly more conducive to higher readability. Though, it is assumed here that the pattern of language behaviour of grade III children in Rajasthan may be indicative to some extent of language pattern of grade III children in other parts of the country "the need to compile state/region-wise norms of children's language pattern cannot be over emphasized."
An analysis of Teachers' views indicates that majority of teachers graded the language textbook as relatively easy. Interestingly their quantitative estimation of percentage of children finding the book as easy/difficult was inconsistent with their own grading of the textbook. In terms of modality of use of textbooks it was interesting to observe that language textbook is the only one which teachers read directly to children and also get the children to read from it in the conventional "Path Pado" (read the lesson) method. Some specific suggestions given by the teachers for improving the readability of language textbook were that the number of lessons in the textbook should be reduced and rearranged in the order of difficulty and composite words and complex sentences should be minimised.
ENVIRONMENTAL STUDIES I AND II TEXTBOOKS From the results of comprehension tests (Table 2) it can be observed that the comparative picture emerging for readability of EVS I textbooks appears similar to that for language textbooks with Karnataka and Haryana textbooks demonstrating lower readability. More than 60 per cent of the students scored less than 50 per cent in these states, demonstrating a negatively skewed distribution. For Maharashtra, Kerala, Tamil Nadu and Assam distributions are positively skewed with 25 per cent to 40 per cent children scoring over 75 per cent.
EVS II textbooks for Tamil Nadu and Maharashtra show high readability (42 per cent to 61 per cent children scoring above 75 per cent). For Haryana and
STUDIES ON DPEP 263
Karnataka the distributions are extremely negatively skewed with 80 per cent to 90 per cent of the sample scoring below 50 per cent indicating a very low level of readability. The textbook of Kerala also does not appear to be pitched at the desired level since over 50 per cent of the sample has scored less than 50 per cent.
Analysis of area-wise differences in readability of EVS I and II textbooks indicate fairly high readability for the urban sample of Assam, urban and rural samples of Tamil Nadu, tribal sample of Kerala and rural and tribal sample of Maharashtra. For the other sub-groups in all the six states readability is towards the lower side. Situation is particularly disturbing for Karnataka where it is low for all the three subsamples. EVS II textbook appears to be generally of a higher difficulty level consistently across all states.
Data obtained for content analysis of EVS I and II textbooks from six states indicate that in terms of "vocabulary" (Table 3) the pattern remains more or less similar to that discussed under language. The points discussed there are thus applicable to EVS I and II as well. Maharashtra and Haryana deviate slightly from this set pattern with percentage of adjectives being higher than that of verbs but evidently this does not in any way affect the readability level of the textbook since both Haryana and Maharashtra textbooks are at two extremes as per the children's performance. However, the relatively higher percentage of adverbs used in EVS I and II textbooks in Haryana could be a factor adversely affecting the readability. This needs further probing. Table 4 gives relative percentages of different types of sentences used in EVS I and II textbooks. Generally a consistent pattern is emerging - maximum use of simple sentences (52.4-91.97) followed by complex sentences (4.76-47.6) and then compound sentences (0.9-19.05). In Tamil Nadu and Kerala textbooks for EVS I and II no compound sentences were reported. In Assam and Karnataka, however, the pattern deviates with the percentage of compound sentences being higher than the complex sentences in EVS I textbook. In EVS II textbook this is seen only in the case of Karnataka. From the data available on sentence type no simple linear relationship seems evident.
Coming to average length of sentences if we consider the data of two extreme states as per children's performance, i.e., Haryana and Maharashtra, the consistently higher average length of sentences is evident for Haryana with compound sentence length of 20 words in comparison to 10.2 in Maharashtra. Sentence length, therefore, could be a potential factor in readability which needs further probing.
EVS I textbook was generally perceived to be easier by teachers across the states. The EVS II textbook was rated as difficult. The modality of textbook use for EVS I and II was different from that of language textbook. Most teachers reported these books being used as guides. They teach from them using their own
264 IER : SPECIAL NUMBER 1995
vocabulary rather than textbook language. Some suggestions given by teachers for improving the texbooks of EVS I and II included - simplification of words, concepts and terms, inclusion of more experiments, maps and illustrations, reduction in number of lessons and quantum of information and increase in exercises for children.
MATHEMATICS TEXTBOOKS As indicated in Table 2 the comparative picture demonstrated a highly negatively skewed distribution for Haryana and Karnataka, 88 per cent and 64 per cent scoring less than 50 per cent. On the other hand, in Maharashtra over 50 per cent of the sample has scored more than 75 per cent marks, followed by Tamil Nadu 47 per cent, Assam 40 per cent and Kerala 31 per cent. The textbooks of the other three states appeared to be high on readability. Maharashtra's mathematics textbooks is almost on the easier side. Area-wise differences reflect mathematics textbook to be readable for all the sub-groups in Assam, for the rural and tribal sub-groups of Maharashtra and Kerala, and urban sample of Tamil Nadu. The picture is most disturbing for Haryana and Karnataka, where mathematics textbook did not cater to linguistic levels and needs of any of the subgroups.
Content analysis as per Vocabulary (Table 3) and Syntax (Table 4) revealed that pattern of vocabulary usage and-sentence type for mathematics is more or less consistent with Language and Environmental Studies I and II. The only deviation for sentence type was in the Karnataka textbooks which used equal number of complex and compound sentences. The only notable aspect that emerges is the wide range across states in the use of compound sentences, 2.09 in Maharashtra (highest readability as per children's performance) to 10.93 in Karnataka, 14.0 Haryana and 19.25 Assam which are towards the low extreme. Greater use of compound sentences would logically lower the readability of the book, depending on children's level of language competence. Coming to the average length of the sentence the data across the state does not indicate any marked deviation which could have a bearing on textbook readability.
As per the teachers comments, mathematics textbook was reported to be difficult in most cases. Modality of use of the mathematics textbook by the teacher was more in the form of a guide. Only 20 per cent of them reported it as being used directly by the children and that too for home assignments.
WRITING AND READING TEST Table 6 shows a consistent trend across the states with the reading scores being higher than the scores on writing. Apparently, the skill of reading is more holistic in nature as compared to writing skill. As observed during the course of test
265
266 IER : SPECIAL NUMBER 1995
administration, even in cases where words were written incorrectly they were read correctly on the basis of the clues provided by the first few alphabets. However, an important point worth noting is that this occurred only in the case of words already existing in the child's own vocabulary. It appears that the familiarity factor facilitates reading. This only further endorses the need for textbook language to be as close to the child's vocabulary as possible. The scores on writing are consistently lower than those for reading, reflecting it to be a more complex skill, involving more of an analytical than synthetic approach.
IS THERE PARITY ACROSS TEXTBOOKS ? Language, EVS I and II and mathematics textbooks would necessarily differ in their structure and approach, since each has a specific objective and scope. While the language textbook primarily aims at development of basic linguistic competencies at the primary level, the EVS and mathematics textbooks are more content oriented. Yet, it is expected that there will be some parity in terms of their difficulty level, particularly with respect to language use. This is necessary if they are prepared keeping in mind the abilities and background of the target group.
When we consider the comparative readability of textbooks for each subject state-wise (Table 7 and 8) variations within subjects are very evident. But the trend of variation is not consistent across the states. In Assam, there is little parity between language and other textbooks in terms of level of language comprehensibility. Significant difference were observed across the subjects excepting EVS I and II with mathematics. In Haryana, while the language and EVS I textbook
STUDIES ON DPEP 267
are more or less at par, the means for EVS-II and mathematics reflect much lower level or readability. Although the situation in Maharashtra as reflected in the mean percentages appears more consistent in terms of high level of readability. Yet, significant differences were observed on mean values across subjects. In Karnataka, while readability of all textbooks appears on the means for the language reader are highest and that for the EVS II textbooks the lowest and these differ significantly from other subjects. Data from Kerala also reflects a fair amount of consistency in terms of level of readability with the means ranging between 51.2 per cent for EVS II and 63.95 per cent for mathematics. Language readability is at par with mathematics and EVS I. Interestingly, Kerala is the only state where mathematics textbooks has the highest means score demonstrating significantly higher readability than EVS I and II. Tamil Nadu again becomes a notable exception with its language reader having the lowest readability (mean 62.4 per cent) and differing significantly from other subjects. This is followed by mathematics textbook (70.35 per cent). EVS I textbook has the highest readability with mean percentage of 76.25 per cent.
The Vocabulary Pattern followed by most states as indicated in Table 3 reflects some minor inconsistencies across the subjects' textbooks. However, there does not seem to be any obvious impact of these in consistencies on readability of textbooks across the subjects.
268 IER : SPECIAL NUMBER 1995
The comparative analysis of sentence usage in different subjects as indicated in Table 4 does not again give any clear-cut trends. However, an interesting observation is that in Tamil Nadu the language textbook which is showing lowest readability as per children's performance has the highest percentage of complex sentences across all textbooks. As a matter of fact for mathematics and language the percentage of complex sentences used in higher than the simple sentences which could logically lower readability. Among the Kerala textbooks, the mathematics textbooks, which demonstrated high readability as per children's performance also has the highest ratio of simple sentences. These observations hint at the possible facilitative influence of greater use of simple sentences enhancing readability. At the same time the lack of any consistent trend indicates the possible interactive/additive effects of various other factors which in combination with sentence type could effect the readability of texts.
Teacher's perception regarding relative difficulty of different textbooks in terms of readability indicate a tendency to rate mathematics and EVS II books as more difficult. Differences across subjects stand out particularly in the context of modality of using textbook by the teachers. While the language book was read by them in majority of cases directly to children and also by children, the EVS I and II textbooks are read by teachers who explain the content in their own words to the children. The mathematics textbook is treated more as a reference book from which problems are taken for demonstration by teachers on the blackboard. The textbooks are used by children only for exercise to be conducted by the children for home assignment.
Investigators' Observations : The overall rating of mathematics and EVS II textbooks as more difficult may be attributed to the inclusion of technical/ complex terminology and high level concepts which may not have been paraphrased in adequately simple language. Consequently, often the teachers themselves are not familiar with the terminology. Even if they are, the complexity of terminology deters them from using these terms in the class. Textbooks for EVS II and mathematics therefore apparently do not lend themselves to varied usage by children without the intervention of teachers. While the teachers intervention can be undoubtedly facilitating, this total dependence on the teacher input may not be very conducive to self learning or learning by discovery/ experimentation on the part of children. It may also create some amount of message dilution depending on the quality of teachers' interpretation of the relevant matter/content. Also, in the context of the existing multigrade situation in schools in the country the need for self-learning instructional material is crucial and the textbook writers should attempt to meet this need. In this context, procuring regular feedback from classroom situations becomes imperative for the writers of textbooks.
STUDIES ON DPEP 269
METHODOLOGICAL ISSUE : APPLICATION OF READABILITY FORMULAE With respect to Application of Readability Formulae it was observed that the Fog Index value as calculated for the language textbook fell much beyond the considered age for Hindi language in the present study and also of Agnihotri and Khanna, 1993; and does not appear to be workable at all for Kannada, Marathi and Malayalam textbooks. This may be due to the polysyllabic nature of these languages which tends to shoot up the index value in a distorted way. The application of readability formulae on Indian textbooks thus needs to be undertaken with caution.
Implications The following implications have emerged from the results and discussion with respect to textbook preparation, teacher training in use of texbooks and methodological aspects of readability measurement.
Textbook Preparation - To ensure continuity and parity in terms of level of readability, there is need to maintain better coordination and interlinkages between different subject textbooks for a particular grade and between textbooks of preceding and succeeding grades.
- To improve readability of the Mathematics and EVS textbooks the possibility of introducing their technical vocabulary alongside, through the language textbooks of the preceding or same grade in a more comprehendable and interesting way, needs to be explored.
- There should be greater involvement and participation of language experts or writers in the preparation of textbooks to ensure parity in terms of language usage.
- Textbook writers of Mathematics and EVS should review their emphasis on use of technical vocabulary right from the early primary level. Interaction with textbook writers gave indication of this emphasis, while teachers in practice, were observed to be using their own termionology which they believed was more comprehendable to children, without exposing them to the given terms.
- Textbook writers should therefore also maintain closer interaction with teachers and conform to a regular cycle of feedback.
- EVS II and mathematics textbooks were generally observed to be low on readability as compared to the language and EVS I textbooks. These should therefore be prepared using simpler language, a more interesting format and
270 IER : SPECIAL NUMBER 1995
more activity-based exercises with a view to enhance their readability and usability.
- The study highlighted differential modes of use of textbooks, with EVS and mathematics textbooks being used more by teachers as reference books or guides. The Language Reader was used directly by children. These textbooks should therefore be reviewed and prepared also with a view to promote self-learning by children, particularly in the context of widely-practised multigrade teaching and high pupil-teacher ratio in schools.
- Use of familiar vocabulary was observed to be a facilitating factor in acquisition of reading skills. In this context, there is a need to compile children's vocabulary for different grades on regional/state basis which could be used as a benchmark for textbook preparation.
- Greater use of complex/compound sentences and lengthy sentences were observed to raise the difficulty level of the text. While these may not be direct predictors of readability, they seem to be crucial in interaction with other factors. In this context, specific guidelines need to be prepared for textbook writers keeping in view the possible impact of language used, particularly in terms of Syntax.
TEACHER TRAINING - During inservice training, interactive sessions between textbook writers and practising teachers should be provided for, which would be of mutual benefit in terms of acquainting teachers with the writer's perceptions regarding modality of use and providing feedback to writers for textbook revisions.
- Training programmes, both preservice and inservice, should also consider the development of sensitivity among teachers towards the need and ways of carrying out effective and ongoing evaluation of the teaching-learning process including the instructional material. In the context, teachers may be familiarized with curricular-based assessment procedures to facilitate evaluation.
METHODOLOGY FOR READABILITY MEASUREMENT - Readability studies should be conducted with a more comprehensive design including aspects like legibility of print, illustrations, idea density etc. for more detailed analysis and recommendations for textbook preparation.
- Thematic content analysis of textbooks is also necessary along with the linguistic analysis for ensuring inclusion of necessary inputs in textbooks for facilitating mastery of minimum levels of learning in the target group.
STUDIES ON DPEP 271
- Readability formulae available in literature should be attempted with caution for Indian Languages since the highly polysyllabic nature of Indian Languages makes the validity of these formulae doubtful.
REFERENCES Agnihotri, R.K. & Khanna, A.L. (1991). The readability of school texbooks : An Indian Study, Journal of Reading, 35:4.
Alliende, F. (1987). Four quick procedures to determine Text Readability. Lectura Y. Vida, 8 (4), 12-22.
Alliende F. (1990). Readability evaluation of written materials. Lectura Y. Vida, 11 (S), 14-18.
Burk, C. (1936). A study of the influence of some factors in style of composition on the interest, comprehension and rate of reading of fourth grade pupils. Journal of Experimental Education, IV, 30-52.
Currie, H. (1990). Making texts more readable. British Journal of Special Education, 17(4) 137139.
Dale, E. & Chall J.G. (1949). The Concept of Readability, Elementary English, XXVI, 19-26.
Engleman, F.E. (1936). The relative merits of two forms of discourse when applied to children's factual content reading materials, Journal of Educational Research, XXIX, 524-531.
Fielding, E. (1990). Reading in the content areas. The importance of choosing appropriate level texts. Reading, 24 (3), 179-184.
Fries, P.H. (1986). Language Features, Textual Cohesion and Reading, Word, 37 (1-2) 13-29.
Fry, E. (1990). A Readability formula for short passages. Journal of Reading, 33/80, 594-597.
Grapstein, R.R. (1985). Sex differences in cognitive functioning, Dissertation Abstracts,Vo. 45, No. 10, 3334-B.
Gross, P.P. & Sadowski, K. (1985). For Index - A, Readability formula program for microcomputers. Journal of Reading, 28(7), 614-618.
Hargis, C.H. (1987). Curriculum based assessment : A Primer spring field, IL:Charles C. Thomas.
Irwin, J.W. (1986). Cohesion and comprehensions. Lectura Y. Vida, 7(4), 4-11.
Lorge, I.C. (1949). Readability Formulae and Evaluation. Elementary English, XXVII, 86-95.
Newton, D.P, (1992). The level of Abstraction of textual materials; a new and an old measure compared. Journal of Research in Reading, 15(2) 117-119.
Peterson, E.M. (1954). Readability in the Social Studies, Bureau of Publications, Teachers College Columbia University, New York.
Shinn, M.R. (Ed.), (1989). Curriculum based measurement: Assessing special children, New York: Guildford Press.
Stone, C.R. (1938). Measures of simplicity and beginning texts in reading. Journal of Educational Research, XXXI, 447-450.
Strang, R. (1948). Principles of Readability applied to reporting research. Teachers College Record, 449-51.
272 IER : SPECIAL NUMBER 1995
Yoakam, G.A. (1948). A Technique for Determining the Difficulty of Reading Material. University of Pittsburgh.
Witting, M.A. & Peterson, A.C. (Ed.) (1979). Few Related Differences in Cognitive Functioning. New York : Academic Press.
Vygotsky, L.S. (1978). Mind in Society. Cambridge Mass Harvard University Press.
Vygotsky, L.K. (1980). The agencies of higher mental functioning. In Wertsch, J.V. (Ed.). The Cancept of Activity in Soviet Psychology, Armonky NY : M.E. Sharpe.
Wilson, M.C. (1948). The effect of simplifying material upon comprehension. Journal of Experimental Education, XIII, 5-8.
273