Enrolling and Retaining Slum Children in Formal Schools

Enrolling and Retaining Slum Children in Formal Schools A Field Survey in Eastern Slums of Kolkata

S

Economic and Political WeeklyJune 2, 20072092the total enrolled students as the population, the selection ofthe schools was done according to the reach out of the NGOvolunteers.IPrimary School Students in the Slumsof East KolkataLocalityCensus 2001 was the first census in India that collected dataon slums5 in several Indian towns. According to this census,the size of the slum population in the Kolkata Municipal Cor-poration (KMC) is 14.91 lakhs which is 32.55 per cent of thetotal population. Slums were identified by Census 2001 in asmany as 107 KMC wards. The contiguous region of eastern slumsof the city, the field area for this study, was located in the Ward56 to Ward 66 of KMC. The demographic profile of the selectedslums (as the Census 2001 data indicate) is presented in Table1.The percentage of slum population varied widely among the11 selected wards. In some of the wards, the percentage of slumpopulation was very high (as high as 99.98 per cent in Tangra,Tiljala) whereas in some other wards the percentage was quitelow (as low as 12.85 per cent in Park St, Taltola, ShakespeareSarani). The proportion of slum children to total child populationin the ward had been more than 80 per cent in five among the11 selected wards. However, there existed a wide inter wardvariation in this regard, as well. For example, the percentage hadbeen as high as 100 in Tangra, Tiljala (Ward 57) and as lowas 16.72 in Park St, Taltola (Ward 62).It is widely believed that the percentage of literate people islow in the slums of east Kolkata. But according to Census 2001,in many of the slums the literacy rate is quite high. In the selectedwards the typical scenario was that the literacy rate was morethan 70 per cent. (In eight out of 11 wards, the literacy rate washigher than 70 per cent. The literacy rate had been found to bemore than 60 per cent in three out of 11 wards of the field area.)In fact, as Table 1 indicates, in most of the selected wards, theliteracy rate in non-slum regions had not been much higher thanwhat it was in the slums of the ward.As the experts observe, literacy among the mothers is crucialfor attracting a child to elementary education. More educatedthe mother is, higher is the value for formal education of children.The conventional wisdom is that slum children remain unedu-cated because the mothers in the slum region are illiterate. ButCensus 2001 indicates that the mothers in the slums of Kolkataare not devoid of education. The female literacy in the slum ofWard 58 (the biggest slum in the city) was as high as 56.03 percent at the time of our survey. In Karya, Tiljala (Ward 64) therate was still higher (66.43 per cent). In Ward 60 the femaleliteracy rate had been 67.77 per cent. In fact, with the exceptionof Tangra, Tiljala, the female literacy rate in the slums of theselected wards had been more than 60 per cent.The physical facilities and socio-economic endowment of theslumdwellers are, however, quite adverse to healthy living. Usually,the residents here live in tiny rooms without proper ventilation.They use common toilets which often remain unclean, theyprepare food in unhygienic conditions and they can never affordto offer a child a quiet atmosphere where the child can prepareher lessons. The sewerage here remains unclean. The factoriesproducing rubber and leather products situated in these areas alsopollute the local sewage. As a result, the children suffer fromseveral chronic diseases like stomach disorders, skin problems,common cold, etc. The other reason for high morbidity is chronicmalnutrition. The level of family income is very low, childrentherefore hardly get proper and sufficient food. As a result, theysuffer from malnutrition and consequently they become suscep-tible to diseases. The incidence of absence due to sickness isquite high in the schools in this locality.Migration is another factor that hampers the schooling of thechildren of the poor families of this locality. The parents of thechildren who are engaged in petty jobs in different factories haveno job security. Periodic job loss is a common phenomenon here.As the family fails to get an alternative means of livelihood inthe locality, they move to some other locality, not necessarilynear the school where the child has been enrolled. Moreover,large portions of the population in these slums are immigrantsfrom Bihar who seasonally migrates to Bihar, particularly at thetime of harvesting. The casualty is the schooling of the child.The SchoolsThe schools in which the survey was conducted account for52.53 per cent of the total schools in the eleven wards of eastKolkata. The major service provider of the selected schools isthe KPSC. It covers 68.52 per cent of the total students for whomthe programme of monitoring the school attendance was launched.This is consistent with the distribution of school facilities in theslums of Kolkata. Almost in every slum, there is at least onegovernment-sponsored primary school run by the Primary SchoolTable 1: Demographic Profile of the Slums in Kolkata (Ward 56 to Ward 66)WardWard Name12345678956Beniapukur83.08358518.863714317585.4970.0767.2363.0357Beniapukur Tangra, Tiljala86.29387569.594194371588.5872.2570.5168.4158Tangra, Tiljala99.988660510.3089228922100.0061.8461.9356.0359Beniapukur, Tapsia68.904592211.256036516785.6068.0666.9663.3660Beniapukur23.64100448.90386589423.1373.5471.0167.7761Beniapukur, Park St24.3182959.36311377624.9380.0873.7368.2862Park St, Taltola17.7681368.87431872216.7277.8477.2675.3263Park St, Taltola, Shakespeare Sarani12.8541829.49180739721.9789.0480.9578.1564Beniapukur, Karaya54.21145727.191865104756.1475.5774.5870.7465Karya, Tiljala91.97738109.157031675696.0972.3970.8366.4366Tiljala72.665137711.967968614377.1073.8970.0863.57Note:(1) Percentage of slum population, (2) Slum population, (3) Percentage of slum child population to total slum population, (4) No of children in theward,(5)No of children in the slum, (6) Slum children/ward children, (7) Percentage literate in the ward, (8) Percentage literate in slums, (9) Female literacyrate in the slums.Source:Census of India 2001, Tables 14 and 15. Provisional Population Totals, Series 20, Paper 2 of 2001.

Economic and Political WeeklyJune 2, 20072093Council of Kolkata. For the parents, the first option is a nearbyKPSC school. There are some Kolkata Municipal Corporation-run primary schools in this area. Some of the slum students getenrolled in these schools. Even the NGO-run schools have spacein this locality. The population pressure being very high, theschools run by the state government and the Kolkata municipalitycannot provide a seat for each child of the locality. The NGO-run schools try to meet this supply gap. The other important factorthat creates a space for NGO-run schools is that these schoolsrun on flexible school schedule that suits some of the families.These schools are often “door step” schools and the parents findit safe to send their girl children to these schools. Commonlyknown as Shikshalayas, these schools are run by trained NGOvolunteers.Average space available per student is found to be better inboth the KMC- and KPSC-run schools. This is due to the factthat these two types of schools get regular financial supports fromthe state government and can, therefore, develop their infrastruc-ture or maintain them regularly (Table 2). On the contrary, mostof the Shikshalayas are run in community halls or covered spacesowned by other institutions. It is not possible for them to offera larger space to each and every student; what they can only doprovide a more caring environment (which is reflected by thelowest average student-teacher ratio in Shikshalayas).With respect to other facilities (such as playground, drinkingwater, toilets, etc) that should be made available to the children,quite expectedly, the scenario is quite bad. Most of the schoolsdo not have supply of drinking water or water for other usages.Only 13 of these schools provide toilets for girl students (thenumber of schools in which the girls are admitted is 100). Thechildren get a playground in only nine out of 104 schools in thislocality. However, the data indicate that the KPSC schools aresomewhat better placed in this regard (Table 3).The StudentsThe survey covered 9,969 school children of the selected 104schools in this locality. Information with respect to 56 childrenhad not been available. The data set, therefore, consists of 9,913students. It appears that the gender ratio of this cohort had beenadverse to the girl students (47.07 per cent). The school-specificdistribution of the gender related data, however, indicates thatthe gender ratio had been most balanced in the Shikshalayaswhere the percentage of girl students had been 50.32. Thepercentage of girl students had been higher in Shikshalayasmostly because, as we pointed out earlier, these were “door-step”schools that can be accessed easily by a girl child.One disquieting feature is that a large number of students inthese schools do not study in the age-specific relevant class. Thisis particularly true if we consider the age distribution of thestudents at the higher classes. Normally, a child is inducted intoschool at the age of four in a KG class. While tracking this childover years, it is normal that the child would be located in classI at the age of five and in the subsequent years in class II, classIII, class IV and class V respectively. Thus, at the age of 10+a child should normally leave the school and pursue her studiesin junior high or high schools. What happens here is that evenif a child gets admitted to a nearby school at the age of fouror five in the age-specific relevant class, the study is interruptedfor various reasons as the child intends to move to higher classes.Often the child fails to cope with the academic requirements forpursuing the curriculum at the higher classes. As a result, eitherthe child drops out or she is found in the school not in the age-specific relevant class, but at a lower class. If we consider theage-specific enrolment in the relevant class as the “net enrol-ment”, the net enrolment rate would be found to be abysmallypoor in these schools6 (Table 5).It is observed the net enrolment is very poor in every school.It appears that only a tiny section of the students study here inthe age-specific relevant class. The scenario is not normal, notasnormal as one finds in the schools in non-slum areas of the city.We would assume that the age-specific relevant class for astudent of age five is either kindergarten or class I. Similarly,the age-specific class for a child of age six would be either Ior II. The overlapping age-specific class is assumed to exist fora child of higher ages as well.Even after relaxing the condition of being in the relevant class,ie, even after relaxing the definition of net enrolment in a class,we observe that the scenario is not promising, particularly forTable 2: Space (Sq Ft)-Student and Teacher-Student Ratioin SchoolSchool TypeAverage SpaceAverage StudentsPer StudentPer TeacherKMC10.6747KPSC8.1634Shikshalaya4.9931Source:Field Survey, November 2003-July 2005.Table 3: Distribution of Schools on the Basis of Availabilityof InfrastructureSchool TypeNumber of Schools withPlay GroundDrinkingWater SupplyElectricityToiletWaterfor GirlsKPSC (34)91720278KMC (5)04312Shikshalaya (65)1269563total (104)2127328413Note:Figures in parentheses describe the number of schools in the givencategory; information not available with respect to electricity facility forone KPSC school.Source:Field Survey, November 2003-July 2005.Table 4: Gender-Specific Distribution of the Studentsin Different Types of SchoolsSchool TypeGenderFemaleMaleNATotalKMC3104322744KPSC29973472136482Shikshalaya1349133262687NA005656Total46565236779969Source:Field Survey, November 2003-July 2005.Table 5: Total Number of Students in Age-Appropriate Classby School TypeSchoolTypeAge-Appropriate Class KGIIIIIIIVVTotalKMC2120016KPSC4734622161162SHIK1111010023Total17854823162191Note: Age 4 – KG, Age 5 – I, Age 6 – II, Age 7 – III, Age 8 – IV and Age 9 – V.Source: Field Survey, November 2003-July 2005.

120 100 80 60 40 20 0

5 6 7 8 9 10 +

AGE

Economic and Political WeeklyJune 2, 20072095student, highest level of education in the family, number of familymembers (adult and minor) and the total monthly income of thefamily, as revealed by the family members. The analysis alsoutilised the information on the child from the school records asunder the type of school (namely, KPSC, KMC and Shikshalayas)in which the ward was studying, gender of the child, his/herreligious identity, age of the student, whether the child wasstudying in an age-appropriate class and the municipal ward towhich his/her family belonged.We have also performed a regression analysis with the per-centage of classes attended by a student, on an average, duringNovember 2003-April 2004 as the explained variable and theset of explanatory variables are given below:X1 =School type; KPSC = 1, 0 otherwiseX2 =Gender; Male = 1, 0 otherwiseX3 =Religion; Hindu = 1, 0 otherwiseX4 =Studying in the proper class; age appropriate class = 1,0 otherwiseX5 =Education of female guardian, those who had completedclass X = 1, 0, otherwiseX6 =Highest education level in the family, education levelabove class X = 1, 0 otherwiseX7=Ward number, Ward 57*, 58* and 65* = 0, 1 otherwise(* wards with percentage of slum population 86.29, 99.98and 91.97, respectively.)X8=Age in complete yearsX9=Monthly per capita income in current pricesD =Dummy for attendance behaviourD = 1, if the attendance is more than 85 per cent= 0, otherwise.The results of the original logistic regression (OLS) afterrunning the data in SPSS 10.0, reveal that the suggested variablescan explain the attendance behaviour of the students at a sat-isfactory level (R2 = 0.654, with F = 45.001). There is a cleardifference in two subsets, ie, the students who were more regularin schools (attending more than 85 per cent of the classes) andthe others, belonging to this set of households.9The regression results indicate that there are three factors whichare statistically significant (at least at 10 per cent level of signi-ficance). These are “type of school”, “the education level offemale guardian” and “per capita monthly income of the family”.The per capita income is found to be a statistically significantfactor, but it works in the reverse direction. Less is the level ofper capita income in the family, more regularly the ward attendsclasses. The result contradicts the conventional wisdom but itseems that the sign of the estimated regression coefficient triesto highlight a truth. The schools here provide food ration forthe students. The students coming from very poor families bankon this food ration. One cannot get the ration unless one attendsthe school. To the poor families, the food ration matters verymuch which is why the child is regularly sent to school. At thehigher income level, the family can exercise the option of notsending the child regularly to school if the alternative engagementof the child is more rewarding – more than what is compensatedby the food ration; the child might give some supportive servicesfor sustaining the family at that level of income. The trade-offin this case would work against going to school regularly – afactor that might have contributed to the negative sign of theestimated coefficient of per capita income as an independentvariable of the model.10The other factor, which is found to be statistically significant,is the level of the education of the female guardian of the child.The regression results indicate that the attendance behaviour ofthe child improves when the female guardian’s education levelincreases more than school leaving level (class X).Again, a student studying in a KPSC school was found to attendclasses more regularly than a student in Shikshalaya does.11 Theattendance was found to increase by more than 9 per cent if thestudent was from a KPSC school. The result is statisticallysignificant even at 5 per cent level of significance. Our inter-pretation is that KPSC schools which have relatively betterinfrastructure might be more attractive to the children. That iswhy the attendance behaviour is better in the KPSC schools.Better infrastructural facilities thus appear to have a favourableeffect on class attendance. The other reason might be that a non-KPSC cohort which consists of students belonging to alternativeformal schools (the Shikshalayas) is a difficult cohort. It consistsof difficult boys and girls who have the tendency to withdrawfromformal schools. It is quite likely that the cohort in the Shikshalayaswould be attending classes less regularly. In fact, this is thechallenge that the alternative formal schools are supposed to face.To conclude, the regression results based on the data collectedfrom 250 households indicate that the per capita income, thehighest education of the female guardian and the type of schoolin which the child is pursuing his/her study are the factors thatmight have some influence on the attendance behaviour of thestudent.IIIRetention ProblemRetention RateOut of 9,969 students registered in these schools in the Session2003-04, only 7,723 were found to continue in these schools inthe next session, ie, only 78.52 per cent students was found tobe retained in the next session (Table 10).At the beginning of the third session (May 2005-July 2005)another 1,602 students were found to be discontinuing (Table11),ie, only 61.75 per cent of the students of the initial cohort werefound to be continuing. The classwise details contained in Tables11and 12, however, indicate that the scenario might not be deplor-able as the aggregate data indicate. Out of 3,792 students whowere missing from the initial cohort of 9,91312 students, 2,182Table 9: The Results of OLS RegressionInterceptX1X2X3X4X5X6X7X8X9D^β00.094**(-) 0.0290.008(-) 0.0060.076***0.0190.0330.003(-) 0.068***0.782*s e13.9555.3813.0113.8854.70714.9496.5513.9961.5320.0083.051R2 = 0.654, F = 45.001*Note:*significant at 0 per cent level, ** significant at 5 per cent level and *** significant at 10 per cent level.Source:Field Survey, November 2003-July 2005.

Economic and Political WeeklyJune 2, 20072096had left the school only after class IV or class V, ie, at the endof primary level schooling. Excluding this group, we find thatthe retention rate in these schools had been 87.51 per cent inthe first session and 90.22 per cent in the second session.13 Therates appeared to be quite promising. In session I, there had been7,056 students in PREP – class II level. This sub-cohort shouldbe in the school even at the beginning of session III. What weobserved was that of the total retainable (7,056) students in thecohort of 9,913, the number of students retained after two suc-cessive sessions was 4,463; ie, only 63.25 per cent of themembers of this cohort.The reality, therefore, is that 36.75 per cent of the studentsin this sub-cohort had been missing from schools only after twosuccessive sessions. If this indicates any trend, the retention ratewas expected to go down further in subsequent sessions.Quality of RetentionRetaining a student in a school indicates that the child has beenplaced in the proper social milieu; he or she has not droppedout from school, neither has the child been engaged in otheractivities inappropriate for a person in the lower age group.However, retention of a child in a school does not contributeto his/her learning process, if the retention is not associated withgetting promoted to higher classes and ultimately passing outfrom primary school. We would consider this aspect of theproblem now.Out of 6,121 students retained in all three sessions, 3,250 hadbeen promoted from session I to session II, 2,866 had beendetained in the same class. There had been five cases in whichdemotion took place. Tables 12 contains the summary informa-tion. The performance of this batch of 6,121 students at the endof session II has been summarised in Table 13. Comparing thefindings of Tables 12 and 13, we observe that the schools couldnot apply “no detention policy” with respect to 46.82 per centstudents in the continuing cohort at the end of session I. Therate was 40.45 per cent in session II.IVRole of InterventionThe Sarva Shiksha Abhiyan aims at 100 per cent enrolmentand retention of children in formal schools. As the field dataindicate, even if enrolment is ensured, it is very difficult toimprove the retention rate in schools that enrol mainly childrenfrom disadvantaged families. A sizeable section of the en-rolled students do not attend classes regularly. Even whenthe children are retained in the formal schools, many of themcannot cope with the curriculum of school education and failto get promoted to higher classes. Such children would graduallydrop out from formal schools and join the labour force at a veryearly age.14What kind of intervention is required to improve the retentionrate? Admittedly, the long-run results can only be achieved if,in future, the socio-economic endowments of the disadvantagedfamilies are improved. But can anything be done under the presentdispensation? As we argue in this section, a better monitoringof the school-going children might produce some positive results,particularly in improving the quality of retention which, inter alia,would induce the children to attend formal schools even whenthe general ambiance is not conducive to valuing elementaryeducation.In east Kolakta slums, such monitoring is being done by oneNGO in the locality15 that provides support to the programmeof retaining the disadvantaged children in the formal school byoffering support services to the schools in the form of trackingthe students and intervening through home visits in case ofdefaulting students. The students in the continuing cohort,particularly the students of Shikshalayas, had been most exposedto such an intervention programme as because they were con-tinuing the studies in the same set of schools where the inter-vention programme was in vogue for three consecutive sessions.Table 10: Distribution of the Students over Class in Session I and Session IIClassSession IISession IKGIIIIIIIVVTotal ContinuingDiscontinuedNATotalPREP221214K G13126374119541622116I111371319524624482910II13467128817592572016III1278109413731991572IV215810616710351202V667783NA05656Total1316177717931573115211277232190569969Source:Field Survey, November 2003-July 2005.Table 11: Distribution of the Students over Class in Session IIand Session IIIClassSession IIISession IIKGIIIIIIIVVTotalDis-TotalContinuingcontinuedKG961152159511421741316I3278576331916201571777II1834390125271701921793III11134267116914641091573IV111711181929601152V22110112Total1023985117315591261120612116027723Source:Field Survey, November 2003-July 2005.Table 12:Performance of Students in Session IClass in Session IPromotedDemotedDetainedTotalPREP22KG59211391731I1263110282292II120834391640III1831251435IV2911Total3250528666121Source:Field Survey, November 2003-July 2005.

Economic and Political WeeklyJune 2, 20072097An analysis of the association between the performance(captured through “getting promoted” and “being detained”)of the students and the degree of intervention by NGO mighthelp us understand whether the NGO intervention had anypositive results.The programme of intervention involves visiting the homesof the children who have failed to attend a critical minimumnumber of classes (more than 50 per cent) in a phase (of oneweek duration) and finding out the reasons behind this incidenceof non-attendance. After ascertaining the reasons,16 the inter-vener is supposed to take a strategy,17 mostly in the form ofadvice, so that the child returns to school and attend the classesregularly in the next phase. Such intervention is usually doneby the community volunteers and a section of the motivatedteachers. It is claimed to have produced positive results, particu-larly when a proper strategy is adopted. The present researcherscollected the relevant data on the “success” and “failure” withrespect to such a programme. The results of the survey mightbe presented first.Field visits revealed that sickness was the major reason (32.06per cent of the cases) behind the irregular attendance of thestudents. As we mentioned earlier, the children here generallysuffer from malnutrition because they do not get adequate andproper food. Consequently, a large number of the childrenperiodically suffer from illness (catching cold, running fever,stomach upset are the common diseases that the children sufferfrom in the slums of east Kolkata.). The next important reason(19.71 per cent) for long absence is seasonal and non-seasonalmigration. This arises out of the fact that a section of the familieshere cannot settle permanently because of the absence of regularjobs. Frequently they move from one place to another in searchof a better job. As a result, a child is uprooted from the school.Sometimes it also happens that the family comes back to the oldlocality and the child is again sent to school, usually to thesame school where the child had been enrolled earlier.Consequently, the child is recorded as being absent for certainperiod in a year.18Lack of motivation (18.07 per cent) is foundto be the next important factor for being absent. Joining thelabourforce did not appear to be a major reason of absence (only0.89 per cent).The community volunteers known as shishumitas,in consul-tation with the teachers and the well-meaning community leaders,adopt an “appropriate” strategy. The field data indicate that sucha strategy often produces desirable results. It was observed thatthe shishumitas performed 4,977 visits to the families of the targetgroup of absentee students and out of 4,977 visits, the adoptedstrategies had been found to be successful for 2,433 times, ie,for 48.88 per cent cases the adopted strategies succeeded inraising the attendance of the child concerned (Table 14).Did such intervention have any effect on improving the qualityof retention? We tried to look into this issue and observed thefollowing:As the number of home visits increases the child is found todevelop seriousness in his/her school prescribed studies andconsequently, he/she succeeds in getting promoted in next higherclass. Outside intervention therefore, is conducive to improvingthe quality of retention in these schools. We would elaborate thiswith the help of the field data.During two successive sessions, the NGO volunteers and theteachers visited 4,435 houses; the number of times the visits weredone had been 9,335. On an average, therefore, the family ofa targeted student was visited more than twice during this period.Counselling had been done, and some positive results did emerge.In order to check whether such visits could improve the qualityof retention in a school we took a sub group of 4,629 students,2,683 of whom succeeded in getting promoted to higherclasses in consecutive two academic sessions and the remain-ing 1,946 had to be detained in the same class during thesame sessions. The records of home visits indicate that 1,572students belonging to this group had been exposed to thestrategy of intervention through home visit by the NGOvolunteers and the teachers. As Table 16 indicates, interven-tion had a positive effect on the academic performance ofthe student. Moreover, as the intensity of intervention (cap-tured in repeated visits) increases, the percentage of the stu-dents getting promoted in successive sessions increases. A non-parametric (x 2) test on the data with the null hypothesis thatthere is no association between getting promoted or failed intwo successive sessions and the degree of intervention standsrejected at 5 per cent level of significance at every level ofintensity of home visit.19The results are quite robust. Had there been no interventionby shishumitas, possibly a larger percentage of students couldhave been detained in the same class. The social reality is thatdemotivation increases as a student from the vulnerable sectionof the society – first generation students in many cases – failto go through the curriculum-based requirements in the formalschools. Demotivation leads to dropping out from formal schools.Parents often work out an immediate trade-off between retainingTable 13: Performance of Students in Session IIClass in Session IIPromotedDemotedDetainedTotalKG18109611142I803327851620II1259523901701III1169282671464IV118371192V0022Total353011524766121Source:Field Survey, November 2003-July 2005.Table 14: Extent of Effectiveness of the StrategiesStrategyFailureSuccessTotalSuccessRate1 Advising for medical consultation19828548359.012 Counselling the parents18091818362750.123 Motivating the community members29319749040.204 Meeting the employer52728.575 Provided other referral services36966.676 Seeking help of school teacher1708425433.077 Others664110738.32Total25442433497748.88Source:Field Survey, November 2003-July 2005.Table 15: Intensity of Home Visit and Quality of RetentionNo of Home VisitTotalPC Promoted*192264.43231971.79315678.8546578.465 and above11087.27Total157269.53Note:*Consecutively in Session I and Session II.Source:Field Survey, November 2003-July 2005.

lli