VIEWS: 5 PAGES: 26 POSTED ON: 6/22/2011 Public Domain
LEARNING ACHIEVEMENT AT THE END OF PRIMARY CYCLE IN DPEP STATES Introduction Achievement surveys are conducted to provide data on quality of education in terms of scores on achievement tests administered to students at a given grade level. Such surveys have been conducted across a number of counties by international organizations such as IEA and OECD in the past. One of the recent surveys of this type was conducted as a part of the Third International Mathematics and Science Study (TIMSS) that was carried out by Boston College and IEA in 45 countries in 1994-95. In India a country -wide achievement survey in Language and Mathematics for the students at the end of their primary cycle was conducted by NCERT in 1991. Such survey was conducted again by NCERT recently, but the report has yet to be released. During the last ten years, the major achievement surveys were those that were conducted in DPEP districts as a part of the DPEP strategy to focus on quality of education. These were conducted to monitor the learning achievement of primary school students. The most recent achievement surveys in the series were the Terminal Assessment Surveys (TAS) that were conducted in DPEP districts to assess the achievement of students towards the end of the District Primary Education Programme. These surveys were conducted mainly to find out whether and to what extent the DPEP objective of improving the quality of education in primary schools was achieved. In DPEP there was considerable emphasis on enhancing achievement level of pupils at the primary stage of education and hence clear quantitative targets were set for their average achievement in language and mathematics. The DPEP target was to “raise the average achievement by at least 25% over the baseline level; to ensure achievement of basic literacy and numeracy competencies in full and minimum 40% achievement in other competencies by all the students completing primary level of education.” In the districts that were covered under DPEP, achievement surveys known as Baseline Assessment Surveys (BAS) in these two subjects at the end of grade I as well as the penultimate grade of primary cycle, were conducted at the time of launching the DPEP in order to provide data on the status of achievement at the beginning of the programme. These surveys were repeated after three years in the form of Mid-term Assessment Surveys (MAS) and again at the end of the programme in the form Terminal Assessment Surveys (TAS). These surveys helped in monitoring the progress made in improving the achievement of students as a result of DPEP interventions. Terminal assessment Survey (TAS) was conducted in 49 DPEP Phase I districts in 2001 and in 86 DPEP Phase II districts in 2003. Actually DPEP Phase I started in 42 districts of 7 states (Assam, Haryana, Karnataka, Kerala, Madhya Pradesh, Maharashtra and Tamil Nadu ) in 1994 as a 7-year programme terminating in 2001 whereas DPEP Phase II started in 76 districts spread over 12 states as a 5-year programme in 1997 terminating in 2003. Since some districts were bifurcated and a new state Chhattisgarh was carved out of Madhya Pradesh, there were 49 Phase I districts belonging to 8 states in which TAS was conducted in 2001 and 86 Phase II districts belonging to 13 states in which TAS was conducted in 2003. The states which had Phase I districts had a few Phase II districts also. The states that had only Phase II districts were Andhra Pradesh, Gujarat, Himachal Pradesh, Orissa, Uttar Pradesh and West Bengal. In general, the DPEP districts were educationally backward districts as the main criterion of selection of districts was low female literacy rate. Those with female literacy rate below the national average qualified for inclusion in the DPEP programme. National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 1 Here in this paper we shall discuss the results of Terminal Assessment Surveys conducted in 49 Phase I and 86 Phase II districts in 2001 and 2003 respectively. We shall confine our discussion to the achievement of students at the end of grade III (in the states in which grade IV is the last primary grade) and at the end of grade IV (in the states in which grade V is the last primary grade), and shall not discuss the results at the end of grade I although under TAS students at the end of grade I were also tested. Our main interest is in finding out how much the students had learnt in language and mathematics subjects by the time they were about to complete their primary education. We shall not discuss the results district-wise, but shall be concerned with the state level picture that emerges on aggregating the district –wise results. Further in this paper we shall not compare the results of TAS with those of BAS and MAS to see whether the expected change in achievement of students has occurred or not as a result of DPEP interventions, since our interest is only in knowing about the present status of students‟ learning as indicated by the Terminal Assessment Surveys of 2001 and 2003. Of course, the findings are not valid for the whole state but only that part of the state in which DPEP was implemented, but the results of TAS provide a good idea of the current level of learning of students completing their primary education in the DPEP states. Although initially only the educationally backward districts were selected for coverage under DPEP, it is expected that after 5 to 7 years of DPEP interventions these districts would have become at par with other districts, if not better than them in respect of students‟ learning achievement. That is why we feel that the results of TAS would be fairly typical of the entire state in which TAS was conducted in their DPEP districts. In particular, in Madhya Pradesh and Uttar Pradesh where the number of DPEP districts is large, the findings should be quite representative of the whole state. Sample design The sampling design for TAS was the same as for BAS and MAS. A stratified two-stage sample of about 50 schools was selected in each district. The schools that qualified for selection were the government and private aided primary schools, middle, and secondary schools that had primary sections. Rural and urban area schools constituted the two strata. Although allocation was proportional, it was envisaged that at least 10 schools in the sample would be urban schools. The first stage unit was „block‟ for selecting rural schools and „city or town‟ for selecting urban schools. The second stage units were schools. If in any selected school there were two or more sections in grade III/ IV, one section was selected at random for testing the students. All the students of that section were tested if their number did not exceed 30, but if their number was more than 30, a stratified random sample of 30 was drawn for testing, with boys and girls constituting the two strata. It was thus ensured that boys and girls were proportionately represented in the sample. In this way the sample size of Grade III/IV students who were tested did not exceed 1500 in any district. Actually, it was much less than 1500 in most of the districts since many schools had less than 30 students in grade III/IV. Tests used in TAS In the states in which grade IV was the highest grade, language and mathematics tests that were suitable for those completing grade III, were used for testing the students who were at the end of their grade III. The language test had 65 items, 30 of which tested word knowledge and 35 tested reading comprehension. In the states in which grade V was the highest grade, the tests aimed at measuring the competencies of grade IV, as these were administered to the students who were at the end of grade IV. In this case, the language test had 70 items, of which 35 tested word knowledge and 35 tested reading comprehension. In both tests the „word knowledge‟ part had items in which a pair of words that were either synonyms or antonyms, was given and the students were required to indicate whether the two words had the same meaning or opposite meaning. Thus, every item had two choices of which one was National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 2 the correct response. In the „reading comprehension‟ part of the test, six unseen passages were given and 5 to 7 questions were asked on each passage to test their comprehension. The questions were all of Multiple Choice type in this part of the test. The mathematics tests tested various competencies of grade III and IV. Each mathematics test had 40 items that tested different competencies of grade III or grade mathematics. All the items were of multiple-choice type in these tests. All the tests were prepared by Ed.CIL‟s Technical Support Group for DPEP after try out in several states. The tests were similar to those used in the all India survey of students‟ achievement at the end of primary level conducted by NCERT in 1991. In all these tests, every item was scored 0 or 1 (0 for wrong answer and 1 for correct answer). The test scores were out of the maximum marks, but for discussion of test results these were converted into percentage scores, that is, scores out of 100. Thus for a student whose test score is 24 out of 40, the transformed score is 60. In the following sections, the mean scores and other statistics are being reported for the transformed scores only. State-wise mean scores in grade III tests The five states in which grade III tests were given to the students who had completed grade III are Assam, Gujarat, Karnataka, Kerala and Maharashtra. Of these states, except Gujarat, all had both Phase I and Phase II districts. Tables 1 and 2 show the pooled means and standard deviations of the test scores in these states for language and mathematics respectively. Also, these tables show the lowest and highest district means and lowest and highest district standard deviations. The lowest district sd is not necessarily for the district that has the lowest mean and the same applies to the highest district sd. The districts in which TAS was conducted in 2003 are covered under Phase II. Table 1. Pooled mean and standard deviation (sd) of scores in language of grade III students in DPEP Phase I and Phase II districts in TAS State Phase No. of Lowest Highest Lowest Highest Pooled results for all districts district district district district districts mean mean sd sd Sample mean sd size Assam I 3 58.6 70.7 16.6 20.0 2170 65.3 19.4 II 6 46.1 48.7 6.0 7.5 3586 47.2 6.6 Gujarat II 3 50.5 66.2 16.4 18.1 2517 60.5 18.2 Karnataka I 4 36.5 64.7 10.7 13.1 3479 52.7 17.0 II 7 40.3 67.1 4.7 19.8 6658 50.7 19.5 Kerala I 3 52.0 52.7 16.9 19.3 4042 52.3 17.8 II 3 55.4 56.7 16.6 17.7 3388 55.8 17.1 Maharashtra I 5 43.7 85.7 7.1 24.2 5291 60.9 23.7 II 4 55.7 64.5 17.3 20.2 3971 60.2 18.9 Note: TAS in Phase I districts was conducted in 2001 and in Phase II districts in 2003. National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 3 Table 2. Pooled mean and standard deviation (sd) of scores in mathematics of grade III students in DPEP Phase I and Phase II districts in TAS State Phase No. of Lowest Highest Lowest Highest Pooled results for all districts district district district district districts mean mean sd sd Sample mean sd size Assam I 3 66.8 67.5 17.9 21.6 2170 67.1 19.9 II 6 45.4 54.9 9.8 21.5 3586 50.7 16.7 Gujarat II 3 49.2 61.6 19.2 23.4 2517 56.4 22.5 Karnataka I 4 28.1 56.8 6.8 8.8 3443 45.9 14.6 II 7 39.1 63.8 4.6 5.0 6655 47.4 11.0 Kerala I 3 39.0 44.0 17.1 18.8 4042 40.8 17.9 II 3 56.4 58.7 17.7 18.9 3387 57.4 18.3 Maharashtra I 5 37.6 96.2 8.5 27.0 5397 57.6 30.3 II 4 45.1 63.4 19.1 25.4 3959 55.7 24.0 Note: TAS in Phase I districts was conducted in 2001 and in Phase II districts in 2003. The mean score in language was close to 60 in only Gujarat and Maharashtra and was highest (65.3) in Phase I districts of Assam. In Karnataka and Kerala, the mean score was in the range of 50 to 56. In Assam, in Phase II districts it was lowest, only 47.2. The standard deviation was between 17 and 20, but in Assam Phase II districts and in Karnataka Phase I districts the standard deviations were quite low, 6.3 and 11.7 respectively. The district- to- district variation in mean scores is quite large in every state except in Kerala and Assam Phase II. The difference between the mean scores of Phase I districts and Phase II districts is not large except in Assam where the mean score of Phase II districts is much lower than that of Phase I districts. The district standard deviations of scores in language vary within a narrow range except in Karnataka Phase II and Maharashtra Phase I districts, where the sd of one district differs widely from that of another. The difference between the standard deviations of Phase I and Phase II districts is not large except in Assam and Maharashtra. In Assam particularly both means and standard deviations of Phase I districts differ so much from those of Phase II districts that one would like to question the very veracity of data, if no satisfactory explanation can be found for the large difference between the attainments of students in the two sets of districts. The mean scores in mathematics are relatively higher in Maharashtra, Gujarat, Phase II districts of Kerala and Phase I districts of Assam. The mean score is lowest (40.8) in Phase I districts of Kerala. The difference between the mean scores of Phase I districts and Phase II districts is quite large in Assam and Kerala, but not in Karnataka and Maharashtra. In Assam, the pattern is similar to that of mean scores in language. It is difficult to explain why the mean score is so much lower in Phase II districts compared to Phase I districts. The district-to district variation in mean score is large in every state except in Assam and Kerala. In particular, in Karnataka, one district had as low mean score as 28.1 and in Maharashtra, one district had the lowest mean score of 37.6. Incidentally, the district of Karnataka that had the lowest mean score in National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 4 mathematics had the lowest mean score in language also. Maharashtra has the distinction of having unbelievably high mean score of 96.2 in mathematics in one of its districts. The standard deviations of scores in mathematics are lowest in Karnataka and highest in Maharashtra. The lowest is 11.0 in Phase II districts of Karnataka and the highest is 30.3 in Phase I districts of Maharashtra. The district standard deviations vary greatly in Phase II districts of Assam and Phase I districts of Maharashtra. In Karnataka, the standard deviations of scores within district are quite low in all the districts; the lowest is 4.6 and the highest is just 8.8. On the other hand, in Maharashtra, it is as high as 27.0 in one of the Phase I districts. Distribution of Scores in the grade III tests of language and mathematics Tables showing percentage distribution of scores (out of 100) in language and mathematics for the total of all DPEP districts in the different states in which TAS was conducted in 2001 and 2002, are given a Table I of Annex I. We find that the distributions vary greatly from state to state. Figures 1 to 2 show the distribution of scores for the total of all the districts of Phase I in language and mathematics respectively and Figures 1 shows similar distributions for the total of all the 23 districts of Phase I. In language, the distribution of scores in Phase I districts differs considerably from that of Phase II districts. In Phase I districts the percentage of students having high scores (say, above 60% or 80%) was much greater than that in Phase II districts. In Mathematics too, the trend is similar. It is difficult to explain the difference in the shape of distribution of Phase I and Phase II districts. It may be due to the difference in the nature of districts included in the two phases. As indicators of overall performance, let us find out the percentage of students who performed well and who performed badly in the tests. Table 3 shows the percentage of students who scored (a) below 30% marks, (b) below 50% marks and percentage of those who scored (c) above 60% marks and (d) above 90% marks. Those achieving below 30% can be considered as poor in achievement who had hardly acquired the competencies that were tested. On the other hand, those scoring over 80% can be considered as having acquired mastery level learning in the subject. The performance of those who scored between 30% and 50% could be considered as fair but not quite satisfactory, while those scoring between 60% and 80% could be treated as good in performance, though not excellent with mastery level learning. Table 3: Percentage of students scoring below 30% and 50% marks and above 60% and 80% marks in grade III tests of language and mathematics in TAS (2001) conducted in Phase districts and in TAS (2003) conducted in Phase II districts Language Mathematics State Phase N Below Below Above Above N Below Below Above Above 30% 50% 60% 80% 30% 50% 60% 80% I (3) 2170 3.18 22.81 57.65 25.53 4268 3.00 22.35 61.52 27.47 Assam II (6) 3608 0.08 69.84 1.16 0.03 2781 12.42 15.85 0.64 0.11 Gujarat II (3) 2517 3.69 32.18 50.18 18.36 4034 15.81 44.54 41.52 15.57 I (4) 3479 16.84 52.49 36.59 13.48 6979 29.99 54.67 33.94 6.74 Karnataka II (7) 6658 19.09 56.19 29.98 9.49 10179 27.63 57.76 31.75 13.63 Kerala I (3) 4042 10.22 47.72 32.41 5.59 10738 28.50 66.95 13.24 1.62 I (5) 5374 11.57 34.96 52.68 27.30 5695 24.40 49.14 43.14 24.16 Maharashtra II (4) 3971 6.22 29.34 50.39 14.53 4007 18.19 42.06 45.24 15.56 Total I (15) 15065 11.22 40.68 44.24 18.03 27680 23.80 51.54 35.46 14.62 National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 5 II (23) 16754 9.63 49.16 31.65 9.98 21001 20.34 43.03 29.70 11.46 Figure 1: Percentage distribution of scores in class III tests in language and mathematics in TAS conducted in 15 Phase I districts of Assam, Karnataka, Kerala and Maharashtra in 2001 and in 23 Phase II districts of Assam, Gujarat, Karnataka and Maharashtra in 2003 Percentage distribution of scores in language Class III in 15 Phase I districts in TAS (2001) 20.00 15.00 10.00 5.00 0.00 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 Percentage distribution of scores in language Class III in 23 Phase II districts in TAS (2003) 30.00 25.00 20.00 15.00 10.00 5.00 0.00 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 6 Percentage distribution of scores in mathematics Class III in 23 Phase II districts in TAS (2003) 30.00 25.00 20.00 15.00 10.00 5.00 0.00 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90- 100 7 Percentage distribution of scores in mathematics Class III in 15 Phase I districts in TAS (2001) In Phase and 20.00 Phase II district, 15.00 11.22% and 10.00 9.63% students 5.00 0.00 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90- 100 respectively had scored below 30% marks in language. The percentage of those scoring over 80% marks was, however, much higher in Phase I districts (18.0%) compared to 10.0% in Phase II districts. While in Assam, very few scored below 30% (only 3.2% in Phase I and less than 1% in Phase II districts, Karnataka the percentage of those scoring below 30% was highest (over 16%) of all the states in both Phase I and Phase II. The percentage of students who had scored over 80% was fairly high (over 25%) in Phase I districts of Assam and Maharashtra and quite low (only 5.6%) in Kerala Phase I. In Phase II districts, the percentage of those scoring over 80% was fairly large in Gujarat and Maharashtra (18.4% and 14.5% respectively) but low (below 10%) in other states. In Mathematics, in Phase I districts, the performance was good only in Assam (only 3.0% scoring below 30% and 27.5% scoring over 80%). In the total of 15 Phase I districts, 23.8% students had scored below 30% whereas in 23 Phase II districts, 20.3% had scored below 30%. However, the percentage of those scoring over 80% was a little higher in Phase I districts (14.6%) compared to 11.5% in Phase II districts. The percentage of those scoring below 30% was between 24 and 30 percent in Karnataka, Kerala and Maharashtra. In Maharashtra, however, the spread and quite wide as 24."% students scored over 80%. In Karnataka and Kerala, very few (6.7% and 1.6% respectively) scored over 80%. To sum up, from the results presented above, can we say that the achievement of students in language and mathematics at the end of grade III is satisfactory? Although we have no yardstick to decide the cut off point below which the achievement can be said to be unsatisfactory, it is clear that the performance of students was not up to the mark in most of the districts where the mean score was, say, below 50. Even with a high mean if the standard deviation is large, it indicates that there were quite a few students who had very low achievement scores whereas some had very high scores. If we consider a mean score of 60 or more as satisfactory performance of the group taking the test, then if we exclude Phase I districts of Assam, only Gujarat and Maharashtra had satisfactory achievement in language and no state had satisfactory achievement in mathematics. In Assam (Phase I districts) the mean score is above 65 in both language and mathematics but on comparing the results of Phase I districts with those of Phase II districts, we find that either the former are way ahead of not only the Phase II districts but most districts of other states too or something went wrong with testing of the students in these districts. 8 State-wise mean scores in grade IV tests Grade IV tests were used in the states in which grade V was the last primary grade. Tables 4 and 5 show the pooled mean scores of language and mathematics test respectively in these states. These tables also show the lowest and highest district means and standard deviations in the same way as Tables 1 and 2 do. The districts in which testing for TAS was done in 2001 are Phase I districts whereas all those in which testing for TAS was done in 2003 have been designated as Phase II districts. Table 4. Pooled mean and standard deviation (sd) of scores in language of grade IV students in DPEP Phase I and Phase II districts in TAS State Phase No. of Lowest Highest Lowest Highest Pooled results for all districts district district district district districts mean mean sd sd Sample mean sd size Andhra II 5 66.2 71.3 4.4 5.6 4403 67.9 5.0 Pradesh Chhattisgarh I 9 27.1 56.0 7.3 21.0 5579 44.0 16.2 II 6 45.0 54.7 17.5 37.2 4033 50.1 27.1 Haryana I 4 44.6 58.6 14.5 19.9 4268 52.2 19.0 II 3 45.1 56.5 16.1 19.3 2781 50.7 18.9 Himachal II 4 44.0 48.8 8.1 9.2 2333 46.0 8.7 Pradesh Madhya I 17 47.1 77.5 12.7 24.6 11002 59.9 16.4 Pradesh II 16 41.8 78.4 13.2 19.3 10190 61.6 18.7 Orissa II 8 35.6 41.5 9.4 12.3 5067 39.9 10.7 Tamil Nadu I 4 67.8 91.9 8.6 19.6 5695 79.0 18.1 II 3 59.5 78.5 15.1 17.7 4057 69.1 18.1 Uttar II 18 55.1 78.9 8.1 19.9 11039 69.5 18.0 Pradesh Note: TAS in Phase I districts was conducted in 2001 and in Phase II districts in 2003. Table 5. Pooled mean and standard deviation (sd) of scores in mathematics of grade IV students in DPEP Phase I and Phase II districts in TAS State Phase No. of Lowest Highest Lowest Highest Pooled results for all districts district district district district districts mean mean sd sd Sample mean sd size 9 Andhra II 5 51.2 58.2 5.3 7.9 4384 54.7 7.0 Pradesh Chhattisgarh I 9 29.4 40.1 6.2 22.1 5579 34.8 14.0 II 6 30.4 39.4 13.2 19.8 4034 36.1 16.9 Haryana I 4 44.8 59.9 14.0 19.9 4268 49.8 18.8 II 3 29.5 49.7 18.8 23.1 2781 40.5 22.2 Himachal II 4 44.2 55.0 16.1 19.5 2338 49.6 17.1 Pradesh Madhya I 17 32.0 70.0 14.0 22.6 10707 52.1 20.5 Pradesh II 16 37.4 75.3 14.7 25.3 10179 54.2 22.6 Orissa II 8 32.9 48.3 14.0 21.1 5067 43.2 18.1 Tamil Nadu I 4 61.3 90.2 13.0 24.7 5695 74.5 22.6 II 3 47.9 74.3 18.5 21.2 4057 61.1 22.6 Uttar II 18 52.8 75.4 10.4 23.5 11039 64.2 18.7 Pradesh Note: TAS in Phase I districts was conducted in 2001 and in Phase II districts in 2003. We find that in the language test the mean score was highest (79.0) in Phase I districts of Tamil Nadu. Actually one district had unbelievably high mean score of 91.9. Next to Tamil Nadu was Uttar Pradesh, in which the mean score was as high as 69.5. Also close to it was Andhra Pradesh with a mean score of 67.9. The states in which the mean scores were rather low (in the range of 40 to 50 roughly) are Chhattisgarh, Haryana, Himachal Pradesh and Orissa. The lowest mean score is 39.9 in Chhattisgarh, whereas the mean score of Haryana is close to 50 in both Phase I and Phase II districts. In Madhya Pradesh the mean score is close to 60 in both Phase I and Phase II districts. In Haryana and Madhya Pradesh there was hardly any difference between the mean scores of Phase I and Phase II districts, but in Tamil Nadu the mean score of Phase I districts was much higher than that of Phase II districts. On the other hand, in Chhattisgarh the mean score was higher in Phase II districts compared to Phase I districts but both were quite low. The variation in district mean scores is quite large in every state except in Andhra Pradesh and Orissa. But the difference between the two is that while in Andhra Pradesh all the districts have high mean scores, in Orissa all have quite low mean scores. As regards variability in students‟ scores, we find that the pooled standard deviations were low (between 5 and 11) in Andhra Pradesh, Himachal Pradesh and Orissa. In Andhra Pradesh, sd was as low as 5.0, which shows that most students scored marks within a narrow range about the mean. The highest sd was 27.1in Chhattisgarh (Phase II). In all other states the sd ranged between 16 and 19. Among the districts, there was one in Chhattisgarh that had as high sd as 37.2. Also, one in Madhya Pradesh had a high sd of 24.6. It will be interesting to study why some districts had very low or very high standard deviation. 10 In mathematics the mean scores are, in general, lower than those of language. The lowest mean score is about 35 in Chhattisgarh (both phases) and the highest is 74.5 in Phase I districts of Tamil Nadu. Uttar Pradesh is next with a mean score of 64.2. The mean scores are also low (between 40 and 50) in Haryana and Orissa. Andhra Pradesh and Madhya Pradesh are a little better with mean scores lying between 50 and 55. The difference between mean scores of Phase I and Phase II districts is small in Chhattisgarh and Madhya Pradesh, but large in Haryana and Tamil Nadu. In both these states, the mean score of Phase I districts is much higher than that of Phase II districts. The district level mean scores differ greatly in every state but the differences are relatively small in Andhra Pradesh and Himachal Pradesh. There are districts in Chhattisgarh, Madhya Pradesh and Orissa in which the mean score in mathematics is between 30 and 33, that is, even below the traditional pass score of 33, whereas there is a district in Tamil Nadu that has mean score of 90.2. The difference between the lowest and highest district means is very large in Madhya Pradesh and Uttar Pradesh, which is understandable because both had a large number of districts under DPEP. But in Tamil Nadu also the districts differ greatly even though very few districts were covered under DPEP in both the phases. Coming to variability in scores, we find that the overall standard deviation was lowest (only 7.0) in Andhra Pradesh, whereas in all other states it was between 14 and 23. Tamil Nadu, Madhya Pradesh and Haryana (Phase II) had high variability in mathematics scores with standard deviation in the range of 22 to 23. The district level sd was lowest (6.2) in one district of Andhra Pradesh and highest (25.3) in one of the Madhya Pradesh districts. The variation between districts in respect of sd was largest in Chhattisgarh and Uttar Pradesh as both had districts with very low and very sd. Distribution of scores in grade IV tests of language and mathematics The distribution of scores in grade IV tests language and mathematics is given in Table II of Annex I for the Phase I and Phase II districts of the different states. Figures 2. In language, the distributions are negatively skewed, indicating a high percentage of students scoring marks the upper range in both Phase I and Phase II districts. In Mathematics the distribution of scores of Phase I districts differs widely from that of Phase II districts. Table 6 shows the percentage of students who scored marks in different broad ranges of marks as in Table 3 for grade III. We find that while 22.8% students were in "over 80%" bracket in Phase I districts, only 11.5% were in this bracket in Phase II districts. But in both cases, about 16% students had scored less than 30% marks. In Phase I districts of Tamil Nadu, the performance was exceptionally good, as over 50% students had scored more than 80% marks and only 5% were in the "below 30%" category. The districts in which the performance was worst were Phase II districts of Chhattisgarh, as nearly 40% students had scored less than 30% marks and only 0.6% had scored over 80% marks. However, in Phase I districts of Chhattisgarh, the picture was different as 31% students had scored over 80% marks. Since in Himachal Pradesh and Orissa only 5.9% and 2.8% respectively had scored over 80% marks; the overall performance was rather unsatisfactory. In Madhya Pradesh (both phases) the performance was a little better as between 18 and 21 percent students scored below 30% and 10 to 12 percent scored over 80%. In Uttar Pradesh also, the overall performance was rather good as only 5.5% students had scored below 30% marks and 20.5% had scored over 80% marks. Table 6: Percentage of students scoring below 30% and 50% marks and above 60% and 80% marks in grade IV tests of language and mathematics in TAS (2001) conducted in Phase districts and in TAS (2003) conducted in Phase II districts State Phase N Language N Mathematics 11 Abov Belo Belo Abov Abov Below Below Above e w w e e 30% 50% 60% 80% 30% 50% 60% 80% Andhra II (5) 4403 0.00 0.66 70.18 0.02 4404 0.45 63.81 2.11 0.00 Pradesh I (4) 2170 11.94 51.25 32.06 9.56 4268 16.00 54.24 26.55 5.58 Haryana II (3) 3608 15.43 65.48 18.52 2.19 2781 33.59 68.90 21.32 7.70 I (9) 3479 20.38 71.41 12.26 1.77 6979 16.81 45.88 39.15 31.02 Chhattisgarh II (6) 2517 12.72 52.53 26.75 3.25 4034 40.53 79.80 9.32 0.57 Himachal II (4) 2333 4.24 68.24 5.53 0.00 2338 11.42 49.32 29.13 5.90 Pradesh I (17) 4042 5.81 32.80 51.89 16.58 10738 20.53 48.28 36.90 9.48 Madhya Pradesh II (16) 6658 4.24 68.24 5.53 0.00 10179 17.84 45.27 40.09 11.72 Orissa II (8) 5067 17.68 86.62 1.99 0.00 5067 27.59 69.51 16.93 2.80 I (9) 5374 0.51 9.83 81.62 56.17 5695 4.95 19.37 71.61 50.64 Tamil Nadu II (3) 3971 3.33 19.05 68.55 38.21 4007 6.79 32.44 53.71 26.23 Uttar Pradesh II (18) 11339 2.41 16.61 71.66 30.02 11039 5.46 24.57 59.36 20.50 I (34) 15065 8.71 38.92 46.79 20.86 27680 15.69 42.64 43.01 22.78 Total II (63) 39896 6.17 35.64 47.62 15.79 43849 15.84 48.44 35.09 11.46 Figure 2: Percentage distribution of scores in class IV tests in language and mathematics in TAS conducted in 34 Phase I districts of Haryana, Chhattisgarh, Madhya Pradesh and Tamil Nadu in 2001 and in 63 Phase II districts of these 4 states plus Andhra Pradesh, Himachal Pradesh, Orissa and Uttar Pradesh in 2003. 12 Percentage distribution of scores in language Class IV in 34 Phase I districts in TAS (2001) 20.00 15.00 10.00 5.00 0.00 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 Percentage distribution of scores in language Class IV in 63 Phase II districts in TAS (2003) 25.00 20.00 15.00 10.00 5.00 0.00 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90- 100 13 Percentage distribution of scores in mathematics Class IV in 34 Phase I districts in TAS (2001) 16.00 14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100 Percentage distribution of scores in mathematics Class IV in 63 Phase II districts in TAS (2003) 20.00 18.00 16.00 14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90- 100 National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 14 To sum up, from the results presented above, we can conclude that the overall achievement level of students who took the grade IV tests in language and mathematics is not very satisfactory in language in Chhattisgarh, Haryana, Himachal Pradesh and Orissa but can be considered as satisfactory in Andhra Pradesh, Madhya Pradesh, Tamil Nadu and Uttar Pradesh where the mean scores exceed 60. The lowest achievement is that of the students in Orissa, Himachal Pradesh and Chhattisgarh whose mean score is below 50. The best achievement is that of the students of Tamil Nadu, Uttar Pradesh and Andhra Pradesh, as their mean scores exceed 67. In mathematics, only in Tamil Nadu and Uttar Pradesh, the mean score exceeded 60. The next are Andhra Pradesha and Madhya Pradesh, in which the mean score was between 50 and 55. In the remaining four states the mean score was below 50. In particular, the achievement was rather poor in Chhattisgarh and Orissa where the mean score was below 45. These two states were lowest in achievement in language too. It is important to identify the reasons of low performance in these states. Test Reliability and competency-wise achievement in language and mathematics in Karnataka and Haryana For Karnataka and Haryana, facility values of the different items of the language and mathematics tests of grades III and IV respectively were available for Phase I districts in which TAS was conducted in 2001. These enabled us to study the achievement of students on different competencies of language and mathematics. These two states were average states in respect of achievement since the mean percentage achievement was 52.7 in language and 45.9 in mathematics in Phase I districts of Karnataka; and 52.2 in language and 49.8 in mathematics in Phase I districts of Haryana. For this study, we used the facility values of items after pooling the data of all the 4 districts in each state. The variation in the mean percentage scores across the districts was large in both the states but by pooling the data of the districts, we could get the picture of the DPEP districts as a whole for the entire state. To give an idea of the large variation over the districts, the district-wise mean percentage scores are shown in Table 7. Table 7: District-wise mean (%) scores in Karantaka and Haryana Phase I districts Karnataka Haryana District Mean % score (class III) District Mean % score (class IV) Language Mathematics Language Mathematics Belgaum 61.8 56.8 Hissar 58.6 47.8 Kolar 42.2 37.3 Jind 56.4 46.6 Mandya 64.7 55.6 Kaithal 48.4 44.8 Raichur 36.5 28.1 Sirsa 44.6 59.9 Total 52.7 45.9 Total 52.2 49.8 The Kuder-Richardson (KR-20) Reliability of the tests was calculated to find out to what extent the tests were reliable in the different districts. This reliability coefficient measures internal consistency of test items. The districts in which mean scores and standard deviations were low, the test reliability was poor and hence the Standard Error of Measurement was large. Table 8 shows the KR-20 Reliability Coefficient of the tests in the four districts of Karnataka and Haryana as well as the reliability derived from the pooled data of all the districts. National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 15 Table 8: KR-20 Reliability Coefficients of language and Mathematics tests in Karnataka and Karnataka – Class III Haryana – Class IV Districts Language Mathematics Language Mathematics Belgaum .818 .286 Hissar .922 .908 Kolar .731 .282 Jind .927 .886 Mandya .773 -.090 Kaithal .909 .758 Raichur .701 .152 Sirsa .853 .744 Haryana in TAS - 2001 Note – Although negative KR (20) is inadmissible, it occurs sometimes when the reliability is very low. The mathematics test of class III has very poor reliability in all the districts of Karnataka. Actually, the standard deviations of the test are quite low in these districts as a result of which the reliability has become very poor. However, when the data of all the districts are pooled up, then treating the sampled schools as one sample representative of DPEP-I districts, we find the reliability is fairly high for all the tests. The relatively high reliability of language tests is due to the number of items in the test being much more than the number of items in the Mathematics tests. In the analysis that follows, use was made of the facility values of test items that were available for phase I districts of Karnataka and Haryana. Since every item was scored 0,1, the Facility Value of an item is simply the percentage of students who answered it correctly. And since each item (or a set of items) measures some competency, it was possible to assess competency-wise achievement from the item analysis results of the pooled data of four districts in each of the two states. Competency–wise achievement in language tests in Karnataka and Haryana The language tests differ from state to state depending on the state language but are similar in form. In every language the grade III test has 30 items of Word Knowledge and 35 of Reading Comprehension and the grade IV test has 35 items of Word Knowledge and 35 of Reading Comprehension. In the Word Knowledge tests every item is a two-choice item. Given a pair of words, the student had to indicate whether the two words have the same meaning or are opposite in meaning. In the Reading Comprehension tests, a few passages are given and on each passage, a few multiple choice type items are given for answering. Some of the questions test simply the understanding of the content; others test the ability of deriving inference or making judgment about the title of the passage or its main theme. Table 9 shows the distribution of their Facility Values of items and their means and standard deviations in the Grade III test of Kannada language used in the Phase I districts of Karnataka in TAS conducted in 2001. The results are for the pooled data of the four Phase I districts. Since the overall mean score is, 51.0 we can consider the findings to be true for an average state in respect of achievement in language at Grade III level. However, the mean score of the Word Knowledge test was 57.8 whereas the mean score of the Reading Comprehension test was only 44.2 (both expressed as percentage of maximum marks). It means the achievement of students in Reading Comprehension is much lower than their achievement in Word Knowledge test. On the average, only 44.2% students could answer a Reading Comprehension item correctly, whereas the corresponding average percentage for Word Knowledge items was 57.8%. National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 16 Table 9. Distribution of item Facility Values and their means and sd's in Word Knolwedge and Reading Comprehension tests of grade III Kannada in Terminal Assessment Survey conducted in Karnataka, 2001 Word Knowledge Reading Comprehension Facility Value Number of items Facility Value Number of items Below 50.0 4 below 20.0 1 50.0 – 59.9 16 20.0 - 29.9 4 60.0 – 69.9 8 30.0 - 39.9 10 70.0 - or more 2 40.0 - 49.9 5 50.0 - 59.9 10 60.0 - or more 2 Total 30 Total 35 Mean FV 57.8 Mean FV 44.2 Lowest FV 45.7 Lowest FV 5.9 Highest FV 80.2 Highest FV 69.9 sd. of FVs 8.0 sd of FVs 14.6 Although the items of both World Knowledge and Reading comprehension tests vary considerably in Facility Value, the variation is greater in the case of Reading Comprehension test. The Word Knowledge item are more homogeneous (sd of Facility Values being only 8.0 compared to 14.6 of Reading Comprehension test). In the Word Knowledge test, whereas there was an item which only 45.7% students could answer correctly, there was another which 80.2% answered correctly. In the case of Reading Comprehension test, the gap between easiest and most difficult item was much wider; there was an item which only 5.9% students could answer correctly, there was another which 69.9% students answered correctly. Also we find that in the Word Knowledge test, there were only 4 items out of 30 which were answered correctly by less than 50% students, in the Reading Comprehensions test, there were 20 items out of 35 that were answered correctly by less than 50% students. For grade IV Table 10 gives the distribution of item Facility Values, mean scores and standard deviation of item means of the two parts of the Hindi test used in the four Phase I districts of Haryana, as a part of TAS conducted in 2001. The results are for the pooled data of the four districts. The table gives the means and standard deviations of the item facility values (after converting them into scores out of 100). As the mean scores are close to 50, we can consider the results as those of an average state. National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 17 Table 10. Distribution and their means and standard deviations of item Facility Values WK and RC tests of grade IV Hindi in Terminal Assessment Survey conducted in Haryana in 2001. Word Knowledge Reading Comprehension Facility Value Number of items Facility Value Number of items 45 – 50 12 30 – 40 4 50 – 55 13 40 – 50 12 55 – 60 7 50 – 60 11 60 - 65 3 60 - 70 8 Total 35 Total 35 Mean FV 52.7 Mean FV 51.4 Lowest FV 45.6 Lowest FV 31.2 Highest FV 31.2 Highest FV 69.9 sd. of FVs 4.7 sd of FVs 10.2 We find that in both the competencies (Word Knowledge and Reading Comprehension) the average achievement is of the same level in terms of mean score. However, the range of facility values of items is wider in the case of Reading Comprehension test. The low sd of the Word Knowledge test indicates that the items are more homogeneous and fewer items would have served the purpose. In the Word Knowledge test, about 35 percent students could not answer even the easiest question correctly. It does not mean that these students failed to give correct answer to other items; some of them would have correctly answered the items with higher difficulty value. There were 23 out 35 questions that were answered correctly by over 50 percent students. It may, however, be noted that even if the students selected the answer at random in each case, about 50 percent would have got the answer correct since each item had only two choices. The mean score of 52.7 is only slightly higher them 50.0. So far as Reading Comprehension is concerned, it is a matter of concern that about 30 percent students could not answer correctly even the simplest question on Reading Comprehension and there were as many as 16 out of 35 questions that could be answered correctly only by less than 50 percent students. Some of the questions that required inference about the theme and judgment based on understanding of the text proved to be more difficult. The sum up, we may conclude that if the Kannada test was appropriate for grade III in Karnataka the students there was had better knowledge of the meanings of words but had difficulty in reading and understanding the content of a written text. On the other hand, in grade IV in Haryana, the students did equally well in both parts of the test, but their overall performance was close to the typical average of 50% achievement level. Also we find that items of World Knowledge test were more homogeneous than those of Reading Comprehension test in both grades III and IV. Competency –wise achievement in Mathematics tests in Karnataka and Haryana The mathematics tests were designed to test several competencies of Grade III and IV. Here we shall discuss the achievement of students on different competencies of Grade III mathematics in Phase I districts of National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 18 Karnataka and of Grade IV mathematics in Phase I districts of Haryana on the basis of TAS conducted in 2001. In each case, the results are derived by pooling up the data of four districts that were covered in DPEP Phase I. Both tests had 40 items. The mean scores were 47.8 and 49.7 and standard deviations of item means (Facility Values) were 15.0 and 16.5 respectively for the Grade III test (in Karnataka) and Grade IV test (in Haryana). The results are thus for an average state in both cases so far as achievement in mathematics at the end of Grade III and Grade IV is concerned. In the case of Grade III test, the easiest item had 67.7 as the Facility Value and the most difficult item had only 6.5 as the Facility Value. In Grade IV test, the Facility Values of the easiest and most difficult items were 87.1 and 17.9 respectively. Grade III Mathematics test in Karnataka, 2001 There was no item in the test which was answered correctly by more than 67.7%, that is, two-third students. There were only 8 out of 40 items which were answered, correctly by over 60% students. In other words, one- third students failed to answer the item correctly that was easiest item of the test. On the other hand, there were 3 items in the test which were answered correctly by less than 10% students. Given below are some results on the performance of students on the different competencies of Mathematics that were tested. (1) Number (10 items) Competencies tested - Finding the place value of a digit in a 4-digit number; identifying the numeral corresponding to written in words; writing a number in expanded form. Finding the smallest of a given set of numbers; or a number that is next to a given number; arranging a set of numbers in ascending order or descending. (Only maximum 4-digit numbers to be used in all these competencies). Findings - Average FV = 52.5 - About 40% students could not identify the numeral for a given number written in words; - Nearly 40% students could not find the place value of a digit in a 4-digit number; - 48% students did not know which number comes next to a given 4-digit number; - 39% did not know which is the smallest of a given set of 4-digit numbers; - 72% did not know whether in a set of four 4-digit numbers, the numbers are in ascending order or not; - Surprisingly 92% students could not write a number like 'six thousand and sixty' correctly in figures, perhaps due to there being no figure in 'hundreds'. - (2) Addition and subtraction (8 items) Competencies tested - Adding two 3 or 4 digit numbers; subtracting a number from a 3 or 4 digit number. Findings - Average FV = 43.5 - About 39% students could not add two 4-digit numbers in which 'carrying' was required ; - 32% could not solve even the simplest questions on addition and substraction, in which 'carrying' or 'borrowing' was not require; - 59% could not answer a more difficult question in which double borrowing was required, such as 600- 58; - 50% could not tell by how much a given 4-digit number is greater than anoher 4-digit number. (3) Multiplication and Division (8 items) Competencies tested - Multiplication of a 3 or 3 digit number by a 1-digit number; division of a 3-digit number by a 1-digit number without remainder; problem sums on the same. National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 19 Findings - Average FV = 47.9 - About 55% students could not multiply a simple 3-digit number like 103 by a 1-digit number; - 59% could not multiply correctly a simple number like 30 by a given 1-digit number; - Nearly 40% students could not divide an even 3-digit number by 2; - 52% failed to get correct answer on dividing a 3-digit number with last digit 5 by 5. - About 52% could solve a problem sum in which division of a 2-digit number 1 digit number (without remainder) was required (e.g. 20 by 4 or 60 by 3) (4) Fractions (3 items) Competencies tested - Identifying the fraction representing the shaded part in a figure; adding fractions that have same denominator; knowing how to write a certain part of a given number as fraction. Findings - Average FV = 47.5 - 53% students did not know how to write 7th part of a given 1-digit number in the form of a fraction; - 58% students did not know what fraction a given rectangular figure is shaded; - 38% could not add simple fractions in which the denominator was the same 1-digit number. - (5) Money related problem (5 items) Competencies tested - Finding the value of a given set of coins of different denomination; problem sums requiring addition or subtraction of amounts given in rupees and paisas; finding the price of several objects when the price of one object is given in paise. Findings - Average FV = 43.3 - About 52% students could not find the total value of a given set of coins of different denominations, while their value did not exceed one rupee. - About 44% students could not solve a problem sum in which two amounts expressed in rupees and paise had to be added ; - about 68% students donot solve a problem requiring subtraction of n amount in rupees and paise from another similar amount. - About 94% could not the price of 12 objects when the price of one was 50 paise. The question became very difficult because of it involved multiplication and conversion of paise into rupees. (6) Measures of length / distance, weight, capacity (3 items) Competencies tested - Adding two lengths given in kms and meters; conversion of units of weight and capacity. Findings - - About 47% students could not add two distances given in kms and meters; - 49% students could not convert kgms into gms.; - 45% students did not know how many millilitres make a litre. (7) Time (1 item) Competency tested - Reading time in a clock. Findings - National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 20 - Surprisingly 91% students could not need time from a clock in which hours were marked but not numbered. Either very few knew how to read a clock or the clock shown in the picture was not clearly marked. (8) Geometry (1 item) Competency tested - Identifying the line segments in a figure. Findings - - Nearly 44% could not tell the number of line segments in a given pentagon. Probably they did not know what 'line segment' meant. - Grade IV Mathematics test (Haryana, 2001) Out of 40 items in the test, there were three which were answered correctly by over 80% students. There were another 9 items which between 60 and 80 percent students attempted correctly. The most difficult items was the one that only 18% students answered correctly. The competencies that were tested and the performance of students on the items testing these competencies are described below. (1) Number (5 items) Competencies tested - Finding place value of a digit in a 6-digit number; writing 6-digit numbers in words and in numeral form; writing numbers in expanded form; Findings - Average FV = 65.5 - 45% students did not know the place value of a particular digit in a 6-digit number (surprisingly only 35% could not answer another similar question correctly). - While 85% students could identify the numeral corresponding to a 6-digit number written in words, only 69% could write a give number in words correctly. - About 53% students did not know the correct expanded form of a given 6-digit number. (2) Addition and subtraction (6 items) Competencies tested - Adding two 4-digit numbers; subtracting a number from a 4-digit number; problem sums on addition and subtraction. Findings - Average FV = 64.9 - Only 13% students could not add two 4-digit numbers correctly. - Only 19% students could not subtract a 4-digit number from another 5-digit number that did not require borrowing, but 30% could not do a similar subtraction sum in which borrowing was required. - Problem on subtraction sums such as 'what to add to a given number to make it equal to another given number' or 'how much larger is one number compared to another' could not be solved by about 55% students. - 37% could not solve a simple problem in which addition of 3 numbers and subtraction of their sum from 50 was required. - (3) Multiplication and Division (10 items) Competencies tested - Multiplying 3 to 4 digit numbers by a 2-digit number; problems sums on multiplication and division including problems solvable by unitary method. Findings - Average FV = 53.4 National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 21 - 37% students did not know that multiplication of any number by zero gives zero. - 36% students could not multiply a 4-digit number by a 2-digit number. - 51% students did not know that in the process of multiplication by a two digit number, of which two numbers, is the product in the first row. Apparently, there are some students who multiply mechanically without understanding the process. - 43% students could not divide a number in thousands (e.g. 40,000) by a number in hundreds (e.g. 200). Apparently, they got confused by the number of zeros in the dividend and divisor. - 62% students could not divide an even 4-digit number with the second digit as zero, by 2. - 31% students could not solve a problem sum in which a number was to be divided by 50 without remainder. - 38% students could not solve a problem in which both division and multiplication by 1-digit number had to be done; another similar problem could not be solved by 50% students, perhaps because of more plausible distractors. - 61% students did not appear to have understanding of terms like divisor, divident, quotient, probably due to lack of clarity in the question. - 57% students could not find the price of 10 object when the price of one was given in rupees and paise. (4) Factors, LCM (3 items) Competencies tested - Knowledge of a prime number and prime factors of a number finding LCM of two numbers, each below 20. Findings - Average FV = 48.3 - 53% students could not identify a prime number out of given numbers all below 20. - 47% students could not find LCM of two numbers, both below 20. - 55% students did not know the prime factors of a given number that was less than 100. Obviously over 50% students had no concept f prime number or prime factors of a number. (5) Fractions and Decimals (7 items) Competencies tested - Understanding fraction and knowledge of simple fraction and equivalent fractions, conversion of simple fractions into decimals. Findings - Average FV = 36.9 - 49% students had no understanding of decimal; they did not know how 1/10 is written in decimals. 59% students could not convert a simple fraction (with 5 as denominator) into decimals. - About 60% students could not identify equivalent fractions or know how to convert a given fraction into a simple fraction. - .76% students could not identify a simple fraction out of a given set of fractions (with denominators not exceeding 10). - 75% students could not identify the fraction representing the shaded part of a figure. - It is clear that very few students acquired understanding fractions and decimals and what equivalent fractions or simple fraction meant. - (6) Measures of length, weight and capacity (6 items) Competencies tested - Problem sums on length, weight and capacity involving addition, subtraction, multiplication or division in which knowledge of units is required. Findings - Average FV = 33.2 - 52% students could not add lengths expressed in kms and meters. National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 22 - 70% students could not calculate the length in which multiplication of 2.5 meters by a 2-digit number was required. - 67% students could not subtract weight expressed in kgms and gms from 10 kgms. - 66% students could not solve a problem in which division a certain weight in kgms by 50 gms was required. - 64% students could not solve a similar problem in which division of a certain quantity in litres by 250 ml was required. - 82% students could not multiply a quantity given in ml. by 10 and express it in litres. (7) Calendar (1 item) Competency tested - Finding the number of days between two dates. Findings - - 73% students could not calculate the number of days between two dates separated by more than 1 month. (8) Geometry (2 items) Competency tested - Knowledge of the name of a geometrical figure; identifying the geometrical figure from its shape. Findings - Average FV = 53.6 - 54% students did not know the name of the closed figure made by three line segments. - 39% students could not identify a square in a given set of four-sided figures. Summary and Conclusion This paper presents the results of further analysis of achievement test data of Terminal Assessment Survey conducted in Phase I and Phase II districts in 2001 and 2003 respectively. The findings show that the achievement is far below the so called mastery level performance in most of the states. If we consider mean achievement score of over 60% as satisfactory, then in grade III overall for DPEP I, Assam, Gujarat and Maharashtra have shown satisfactory performance in language, while in mathematics only in Assam overall mean exceeded 60%. In Kerala and Karnataka, the mean score remained below 60% in language and below 50% in mathematics. In the case of DPEP II districts of these states, only in Gujarat and Maharashtra the mean scores in language exceeded 60% in all other states, it was below 60%. In mathematics, the mean score was below 60% in all the states so far as DPEP II is concerned. In grade IV, overall in Phase I districts, the language mean score exceeded 60% only in Tamil Nadu and was below 60% in Chhattisgarh, Madhya Pradesh and Haryana. In Phase II districts, the language mean score exceeded 60% in Andhra Pradesh, Madhya Pradesh, Tamil Nadu and Uttar Pradesh and Uttar Pradesh, and remained much below 60% in Chhattisgarh, Haryana, Himachal Pradesh and Orissa. In mathematics, the mean achievement was much below 60% in all the states for both Phase I and Phase II districts, except Tamil Nadu and Uttar Pradesh where the mean score exceeded 60%. The percentage of students scoring over 80% marks in the language tests of grade III was very low (below 10%) in Karnataka (Phase II), Assam (Phase II) and Kerala. In grade IV, less than 10% students scored marks over 80% in language tests in Haryana (both phases), Andhra Pradesh, Himachal Pradesh, Madhya Pradesh (Phase II), Chhattisgarh (both phases), and Orissa. In mathematics, less than 10% students scored marks over 80% in Haryana (both Phases), Madhya Pradesh (Phase I), Andhra Pradesh, Chhattisgarh (Phase II), Himachal Pradesh and Orissa. Only in Tamil Nadu, over 50% students scored over 80% marks in both language and mathematics tests. National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 23 Overall, among the total students of grade III in Phase I districts tested in 2001, 18.0% students scored over 80% marks in language and 14.6% scored over 80% marks in mathematics whereas the corresponding percentages were 10.0% and 11.5% respectively in Phase II districts where testing was done in 2003. The findings are similar for grade IV, where in Phase I districts, 20.9% students scored over 80% marks in language and 22.8% scored over 80% in mathematics. The corresponding percentages were less in Phase II districts where testing was done in 2003, as 15.8% students scored marks above 80% in language and 11.5% scored marks above 80% in mathematics. It will be of interest to analyse the factors responsible for good performance of students in Gujarat, Maharashtra, Uttar Pradesh and Tamil Nadu and for relatively poor performance in Haryana, Orissa, Himachal Pradesh, Karnataka, Chhattisgarh, Madhya Pradesh and Kerala. In Assam, the wide difference between the results of Phase I and Phase II districts need to be investigated. The exceptionally good performance of students in Tamil Nadu also requires study of factors responsible for such performance. Analysis of competency-wise achievement scores in language and mathematics tests in Karnataka and Haryana shows the competencies in these subjects in which the performance is good or poor. In mathematics particularly, the competencies in which the performance is poor are clearly identified. Such analysis needs to be done for other states too in order to identify the hard spots in Mathematics that require special attention in teaching. National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 24 Annex I Table I: Percentage distribution of scores in different states Phase II (TAS, 2002) (A) Percentage distribution of scores in language of grade III students No. of No. of 90- State district students 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 100 Assam 6 3608 0.00 0.00 0.08 14.83 54.93 28.99 0.91 0.22 0.03 0.00 Gujarat 3 2517 0.16 0.91 2.62 9.81 18.67 17.64 18.83 12.99 13.75 4.61 Karnataka 7 6658 3.77 4.24 11.08 17.27 19.83 13.83 12.80 7.69 7.10 2.39 Maharashtra 4 3971 0.86 1.89 3.48 9.27 13.85 20.27 17.12 18.74 10.07 4.46 Total 20 16754 1.72 2.27 5.64 13.73 25.80 19.20 12.17 9.50 7.28 2.70 (B) Percentage distribution of scores in mathematics of grade III students No. of No. of 90- State district students 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 100 Assam 6 3608 12.42 0.00 0.00 1.19 2.25 83.51 0.50 0.03 0.03 0.08 Gujarat 3 2517 2.74 3.38 9.69 12.44 16.29 13.95 12.51 13.43 10.57 5.01 Karnataka 7 6655 5.03 7.71 14.89 15.76 14.37 10.49 9.17 8.96 9.32 4.31 Maharashtra 4 3959 3.81 4.93 9.45 11.01 12.86 12.71 14.37 15.31 10.05 5.51 Total 20 16739 5.99 4.74 9.61 11.00 11.69 27.27 9.03 9.21 7.68 3.79 Table II: Percentage distribution of scores in different states Phase I (TAS, 2001) (A) Percentage distribution of scores in language of grade III students No. of No. of 90- State district students 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 100 Assam 3 2170 0.51 0.32 2.35 7.47 12.17 19.54 16.13 15.99 13.96 11.57 Karnataka 4 3479 2.10 3.13 11.61 16.96 18.68 10.92 13.22 9.89 11.61 1.87 Kerala 3 4042 1.24 2.33 6.66 12.49 25.01 19.87 18.13 8.68 5.00 0.59 Maharashtra 5 5374 1.15 4.06 6.36 11.98 11.41 12.36 11.50 13.88 16.65 10.64 Total 15 15065 1.30 2.84 7.08 12.62 16.85 15.07 14.34 11.87 11.97 6.05 (B) Percentage distribution of scores in mathematics of grade III students No. of No. of 90- State district students 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 100 Assam 3 2170 0.55 0.55 1.89 6.59 12.76 16.13 16.59 17.47 15.81 11.66 Karnataka 4 3444 7.81 8.54 13.65 13.24 11.44 11.38 14.17 13.04 6.71 0.03 Kerala 3 4312 3.62 8.21 16.67 20.18 18.27 19.81 8.46 3.15 1.35 0.28 Maharashtra 5 5398 5.74 7.60 11.02 11.80 9.21 8.30 9.11 11.26 17.01 8.95 Total 15 15324 4.87 6.98 11.91 13.74 12.76 13.34 11.13 10.26 10.11 4.89 National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 25 (A) Percentage distribution of scores in language of grade IV students No. of No. of 90- State district students 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 100 Haryana 4 4268 0.40 2.13 9.23 19.31 20.90 16.45 13.36 8.81 6.35 3.07 Chhattisgarh 9 5579 1.09 7.55 11.74 15.54 35.49 16.33 6.76 3.73 1.67 0.11 Madhya Pradesh 17 10736 0.35 1.54 4.08 11.54 16.21 15.73 17.79 15.71 12.25 4.80 Tamil Nadu 9 5695 0.14 0.09 0.28 2.60 6.73 8.55 11.19 14.26 21.47 34.70 Total 39 26278 0.47 2.60 5.72 11.71 19.01 14.42 13.30 11.73 11.04 10.00 (B) Percentage distribution of scores in mathematics of grade IV students No. of No. of 90- State district students 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 100 Haryana 4 4268 1.59 4.83 9.58 18.11 20.13 19.21 12.51 8.46 4.12 1.45 Chhattisgarh 9 6979 1.15 5.77 9.89 12.57 16.51 14.97 5.69 2.44 10.75 20.28 Madhya Pradesh 17 10738 1.37 5.57 13.60 13.82 13.92 14.83 14.59 12.82 7.66 1.82 Tamil Nadu 9 5695 0.39 1.12 3.44 5.44 8.97 9.03 8.81 12.15 19.77 30.87 Total 39 27680 1.15 4.59 9.95 12.44 14.51 14.35 10.84 9.39 10.39 12.39 National Conference on Enhancing Learning in Elementary Schools , Bangalore, 23-25 July 2004 26