ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Revisiting Dropouts

While progress in improving literacy in India has been remarkable, the phenomenon of school dropouts has remained a blot in the face of an otherwise commendable performance. Dropout rates have undoubtedly come down but are still high enough for us to sit up and take notice. Sex differentials have also reduced. But rates for females have consistently remained above those of males. Since dropping out is a worldwide phenomenon, the issue has been the subject of intense analysis and factors influencing such an outcome have been widely discussed. However, not much is known on how these factors play themselves in the north-eastern part of our country. Additionally, two fresh issues are also examined, viz, familial duties and parental bonding, insofar as their influence on school discontinuance is concerned.

Revisiting Dropouts

Old Issues, Fresh Perspectives

While progress in improving literacy in India has been remarkable, the phenomenon of school dropouts has remained a blot in the face of an otherwise commendable performance. Dropout rates have undoubtedly come down but are still high enough for us to sit up and take notice. Sex differentials have also reduced. But rates for females have consistently remained above those of males. Since dropping out is a worldwide phenomenon, the issue has been the subject of intense analysis and factors influencing such an outcome have been widely discussed. However, not much is known on how these factors play themselves in the north-eastern part of our country. Additionally, two fresh issues are also examined, viz, familial duties and parental bonding, insofar as their influence on school discontinuance is concerned.


I Introduction

t the 1990 World Conference on Education for All, governments agreed to a broad range of education goals including that of attainment of Universal Primary Education (UPE) by the year 2000. Sadly the millennium year had come and gone but the UPE goal is still a distant dream more so in developing countries like India. The millennium development goals as drawn up by the UN now directs nations to ensure that all boys and girls complete a full course in primary education by the year 2015. While the government has been making concerted efforts aimed at expanding the reach of education, the phenomenon of school dropouts remains a blot on the progress of education in India. What is cause for particular concern is the enormity of the problem in all states of the Indian union and at all stages of school education where unacceptably high dropout rates have been reported. Statistics are worring as Table 1 suggests.

Clearly considerable progress has been achieved in the last 40 odd years notwithstanding the occasional hiccup. The year 1992-93 was perhaps a particularly bad year with all stages of education showing increased dropout rates across the board. However the heartening fact is the significant progress made in the reduction of female dropouts. Sex differentials in dropout rates have also reduced considerably over the years. Still then out of every 100 students enrolled in class I, 39 per cent discontinued schooling and dropped out in the primary stage while two-thirds did not go on to complete class X (dropout rate for 01-02) – a worrying situation indeed. Another disturbing angle of dropout statistics is the fact that dropout rates have consistently remained higher for girls as compared to boys.

Why is dropping out a phenomenon to sit up and take notice? Why has it remained and still is a black spot in our educational progress? Human capital theory proposes labour productivity to be a direct function of the amount of schooling received. Higher schooling is expected to increase cognitive development and also contribute to economic modernisation [Colclough 1982]. At the individual level, a student who does not complete school education severely restricts his adult earning potential. In any social interaction – be it operating a bank account, setting up of an enterprise or taking advantage of government schemes, these people find themselves terribly disadvantaged. Simply put, they lose out in the race for self advancement being unable to leverage the opportunities available to lead a healthy and contended life. Besides dropouts cost the nation money. Dropouts are less likely to find and hold jobs that pay enough to keep them off public assistance. Even if they find a job, dropouts earn substantially less than high school graduates. Higher rates of unemployment and lower earnings cost the nation lost production and reduced tax income [Rumberger 2001]. Besides there also exists a large body of evidence which suggests that a significant proportion of our wayward citizens are in fact school dropouts. Two burning social issues in the north-eastern part of India are militancy and drug abuse. Newspaper columns have consistently reported that these militants as well as drug peddlers and their victims are actually school dropouts in a vast majority of cases. This is more so true in the case of lower level functionaries. Besides dropouts engage in crime, have poorer health, have lower rates of intergenerational mobility and lower rates of political participation [Rumberger 1983]. A female dropout is a cause for greater concern in a country like India which has the problem of burgeoning population. The positive association between female education and decline in fertility, infant and child mortality has been well established [Kingdon 1999; World Bank 1997].

Table 1: Dropout Rates at Primary, Elementaryand Secondary Stages

1960-61 1970-71 1980-81 1990-91 1992-93 1999 2001
2000* 02*
Classes I-V
Boys Girls 61.7 70.9 64.5 70.9 56.2 62.5 40.1 46.0 43.8 46.7 38.7 42.3 38.4 39.9
Total 64.9 67.0 58.7 42.6 45.0 40.3 39.0
Classes I-VIII
Boys 75.0 74.6 68.0 59.1 58.2 52.0 52.9
Girls 85.0 83.4 79.4 65.1 65.2 58.0 56.9
Total 78.3 77.9 72.7 60.9 61.1 54.5 54.6
Classes I-X
Boys NA NA 79.8 67.5 70.0 66.6 64.2
Girls NA NA 86.6 76.9 77.3 70.6 68.6
Total NA NA 82.5 71.3 72.9 68.3 66.0

Note: * Provisional. Source: Ministry of Human Resource Development, Government of India.

Economic and Political Weekly December 23, 2006

With dropout reduction now a worldwide concern, it has become necessary to address the question of why students dropout. Do dropouts have certain features which characterise them as distinct from students? Any policy-maker would be interested in answers to these questions as appropriate remedial policy measures can then be designed. Unfortunately, the issues are complex and straightforward all-encompassing answers are still not available. Suffice it to say that a number of studies have been taken up and certain broad patterns have emerged. It is now widely acknowledged that reason for discontinuing school can broadly be classified into three groups [Weber 1989; Rumberger 2001]:

(i) Family-related reasons (socio-economic status, disadvantaged groups, parental education and single parent families); (ii) schoolrelated reasons (attendance, grades, academic achievement, interest in school and school work); and (iii) personal reasons (disciplinary problems and other extenuating circumstances like marriage, etc).

The evidence from India specific studies has not been much different. Ramchandran and Saihjee (2002), in a study of District Primary Education Programme (DPEP) has found that “general household characteristics like income, caste, occupation and educational level of parents continue to determine access, attendance, completion and learning achievements”. Sengupta and Guha (2002) in a study of girl children in the state of West Bengal concluded that the strongest factors with regard to school participation, enrolment and dropout were household factors such as parental schooling, household income and father’s occupation. They also found that caste and religion were important determinants of schooling. Similar results have also been reported by Vaid (2004). She however found only partial and weak support for the effect of caste with regard to school continuance.

II Present Efforts

Much is already known about why students’ dropout – in particular, about the significant part played by family related factors. In a different way, this paper revisits the issue of family related factors insofar as its contribution to dropping out is concerned. Some fresh angles of investigation will be persued. The setting is the largest urban city in the north-eastern part of India and its surrounding areas. I have already mentioned the usual family related factors, viz, socio-economic status, disadvantaged groups which have all been well researched and their effects on the phenomenon of dropping out well documented. In the Indian context, there has been some additional evidence on the existence of a relationship between familial duties and dropping out of school. Familial duties are defined to include household chores as well as work done with parents in outdoor activities. These duties are non-remunerative in nature.

The NSSO in its 50th round (July 1993-June 1994) collected data on the activity profile of children belonging to age group 5-14 years. Information was collected on school attendance, participation in household chores and work-related activities by children in this age group. Results of this survey were published in the NSSO report no 412, in May 1997. Among other highlights this report states that the proportion of children helping in household chores was highest among dropouts and that the sex differential in these proportions were also highest among dropouts; Table 2 provides some details.

As this table suggests, 40.8 per cent (28 per cent) of urban female (male) dropouts helped in household chores. These percentages are significantly higher than the corresponding percentages for the other two category of children (currently attending and never attended) studied. A similar story holds good for rural children. In view of this evidence, one hypothesis to be tested in this paper is whether the burden of familial duties induces a child to dropout. Familial duties are defined to include household chores as well as other duties not directly related to the household. As is well known, a significant activity for children in India particularly those belonging to lower economic strata relates to working with or helping parents in their work.

A second hypothesis taken up for examination is whether the depth of parental bonding is in any way related to the decision to dropout of school. A wide range of empirical research has shown that an unstable mother – child attachment has a negative influence on the childs’s social development [Zsolnai 2002]. Empirical evidence also exists suggesting that appropriate parental style and involvement are associated with school adaptation [Steinberg et al 1994, Dornbusch et al 1987]. In the development of social competence, the factors of import include the nature of attachment between mother and child [Schneider 1993]. It is therefore worthwhile to examine this issue in the specific context of school continuance. In the Indian context, this aspect appears to have been inadequately researched. In many families particularly those in urban areas, both parents work. This means parents have many demands on their time. Working parents are often faced with trying to complete all household duties within the limited time available. The immediate casualty is bonding with children as parents cannot find enough time for them. This problem is perhaps more acute in families belonging to lower economic strata especially in urban areas who are often forced to work long hours outside their homes.

To the best of my knowledge, the possible influence of the two issues outlined above on school continuance has not been examined by any researcher. Testing the validity of the above hypotheses after controlling for the well known factors influencing disengagement from school is the primary aim of this paper. Additionally, an equally important aim is to evaluate and quantify how these well known factors play their part in school discontinuance insofar as the situation in the largest city of northeastern region of India and its suburbs is concerned.

III Evidence from Data

Most studies on dropouts in India have been based upon large administrative surveys which typically provide limited information on individual dropout characteristics, their ways of life and conduct. An attempt has therefore been made to correct these lacunae. Data for the paper has been taken from a dropout focused survey conducted by the department of statistics, Gauhati University, India during 2001-03 in and around the city of Guwahati (sponsored by the ministry of statistics and programme implementation, government of India). With a population of

Table 2: Number of Children of Age Group 5-14 Years Helpingin Household Chores Per 1,000 Children

Status of School Attendance Boys Girls
Rural: Currently attending 260 293
Dropped out Never attended 284 140 404 201
Urban: Currently attending 233 292
Dropped out Never attended 280 106 408 221

Note: Reproduced from Table 4.6 of NSSO report no 412.

Economic and Political Weekly December 23, 2006

8,08,021 (as per the 2001 Census), it is not only the capital of the state of Assam but is also largest city in the north-east. The city has 60 municipal wards. For the survey, one government or government-aided school (henceforth referred to as G and GA school) were randomly selected from each of these wards. In each of these schools, dropouts during the last three years were located and individual data collected about them. The definition of dropouts adopted for the sample survey was similar to the one used by NSSO. Any child who discontinued studies before completing the class they had last attended was classified as a dropout. During data collection, considerable stress was placed upon creating a profile of the dropout prior to the time around which the student disengaged from school. Data was also collected from parents of two randomly selected students of each class of each of the selected schools so as to aid comparison. Additionally, three G and GA schools were randomly selected from each of the five rural blocks around the city and individual data also collected from parents of dropouts and students of these schools. In all, 75 schools were approached and data collected on 983 dropouts and 382 students. The choice of G and GA schools was dictated by two considerations. First, some 98 per cent of educational institutions of Assam are directly or indirectly supported with funds from state budget (budget speech by finance minister Assam for FY2004-05). Second, private schools (the remaining 2 per cent) are almost exclusively patronised by the well-off sections of the society. The problem of dropping out is less acute here. Besides, private schools are reluctant to part with dropout related information for fear of negative publicity.

A few descriptive statistics on the sample will not be out of place here. 50.3 per cent of dropouts and 51.2 per cent of students were females. The distribution of last class attended by male (female) dropouts reveals that while 52.5 per cent (52 per cent) dropped out in the primary stage of education comprising of class I-V, 25.5 per cent (25.6 per cent) dropped out in middle school consisting of class VI-VIII and 22 per cent (22.4 per cent) dropped out in classes IX-X which is the high school. The break-up of dropouts and students included in the sample according to religion and mother tongue shows a diverse distribution. The Assamese are the dominant community of the state of Assam followed by Bengalis. Assamese language is also the official language of the state. Bodo, Rabha, Mising are the language of tribes who reside in this state. Though they have a distinct lifestyle of their own, these groups are often seen as parts of the larger Assamese community.

The analysis of who first suggested to the student to dropout is interesting as Table 5 suggests.

Clearly dropping out is more often than not a self-induced decision. This is consistent with the finding that 84.1 per cent of parents of dropouts surveyed answered in the affirmative when asked if completing school education could have meant better future for their wards. It appears that the positive influence of education has impressed them. This is not surprising as the area of the study is an urban city and its surrounding rural areas well connected to the city. Another noticeable feature is the fact that only 3 per cent of dropouts of the study stated that remunerative work was a reason for dropping out of school. This is consistent with Table 4.12 of NSSO 50th round report no 412 where only

1.3 per cent - 3.2 per cent of children not attending school stated that they had done so for wage or salary.

Even though dropping out appears to be a largely self-induced decision, this is not to suggest that dropping out is an on/off decision. Empirical longitudinal studies have shown that dropping out is a process that begins early in development and continues through to the time a student formally withdraws from school. The process of dropping out perhaps begins prior to the child entering school. The child’s early home environment and quality of early care giving emerged as powerful predictors of whether a student remained or dropped out of high school [Jimerson et al 2002]. Even though dropping out is a long drawn out process, there comes a time in the life of a would-be dropout when the question of discontinuing school props up for the first time. The above table tries to capture the picture of who propped up the question for the first time.

Each dropout/student included in the study was classified according to their standard of living based upon household assets. Three standards were defined, viz, high, medium and low. This classification was done on the basis on family assets and following the methodology adopted in ‘Rapid Household Survey, RCH Project Phase II (1999)’ of the government of India. Table 6 provides break-up of respondents according to standard of living.

For the purpose of this paper, respondents from families with standard of living described as “high” were excluded from analysis. This was thought necessary as their problems are different. Secondly, the problem of dropouts is more concentrated in families with standard of living “low” or “medium” as Table 6

Table 3: Distribution of Mother Tongue in the Sample

Mother Tongue Student Dropout

Assamese 240(62.8) 552(56.2) Bengali 92(24.1) 204(20.8) Bodo 6(1.6) 36(3.7) Hindi 20(5.2) 67(6.8) Rabha and Mising 2(.6) 7(.7) Other languages 22(5.8) 117(11.9) Total 382 983

Note: Figures in brackets indicate per cent within dropouts/students.

Table 4: Distribution of Religion in the Sample

Religion Student Dropout

Hindu 327(86.1) 776(79.1) Muslim 51(13.4) 194(19.8) Christian 1(.3) 6(.6) Sikh 1(.3) 5(.5) Missing data 2 2 Total 382 983

Note: Figures in brackets indicate per cent within dropouts/students.

Table 5: Who Initially Proposed Dropping Out to the Dropout

Frequency Per Cent

Father 29 3.0 Mother 45 4.6 Both parents together 144 14.6 The dropout himself/herself 633 64.4 All together 94 9.6 Others 9 .9 Total 954 97.2 Missing data 29 2.8 Total 983 100.0

Table 6: Classification of Standard of Living of Dropouts/Students

Dropout Student Total

Standard of living Low 611 117 728 Medium 284 184 468 High 24 30 54

Missing data 64 51 115 Total 983 382 1365

Economic and Political Weekly December 23, 2006 suggests. Table 4.4 of NSSO report no 412 relating to 50th round also presents the same picture in the context of India as a whole. It was pointed out in the said report that the number of dropouts per 1,000 distribution of children in the age group 5-14 declined steadily as monthly per capita expenditure (MPCE) increased. One may note here that even in families classified with “medium standard of living”, the assets (and hence incomes) are not high. I have included two classes for standard of living as there is also evidence to the effect that educational attainment is influenced by economic equality even when average incomes are low [Sarin 2003].

IV Analysis

Variables, Their Means and Standard Deviations

The variable of interest is binary denoting whether the respondent is a student or a dropout. The endeavour is to analyse the odds of a respondent being a dropout. Consequently, logistic regression technique is adopted for analysis. Following explanatory variables are thus considered:

  • (i) “Number of siblings” (Tot_sib) of the dropout/student.
  • (ii) “Father’s education” (Father_e) and “Mother’s education” (Mother_e): Both these variables indicate last class attended by father/mother.
  • (iii) “Caste of the dropout/student” (Caste): Scheduled castes and scheduled tribes are socially disadvantaged sections of the society. This variable is coded to take 1 if the dropout/student is from these backward classes and 2 otherwise.

  • (iv) “Bringing water for family use” (B_wat) and “working with parents in outdoor activities” (W_par): These variables are included to understand how the burden of familial duties influence the likelihood of a student to dropout. Both these variables are coded to take 0 if the job (bringing water/working with parents) was not done or rarely done, 1 if done sometimes and 2 if done regularly or almost regularly. Choice of these two variables to reflect the burden of familial duties is led by the fact that they constitute major household chores particularly among the not so well off sections of the society which is my target group.
  • (v) “Taking them out” (TTO) and “Story telling and gossiping” (STG): In some sense, these variables measure the extent of time parents directly spend with their children. The endeavour is to examine if spending time with children reduces the likelihood of discontinuing studies. TTO is coded as 0 if done rarely or never and 1 if done regularly or sometimes. STG is coded as 0 if the activity is not done or rarely done, 1 if done sometimes and 2 if done regularly or almost regularly.
  • (vi) “Religion”: The reduced data set of families in low and middle standard of living contained those professing Hindu, Muslim, and Sikh faith. I clubbed Hindus and Sikhs in one group keeping in view of present day realities. This variable takes 1 if the dropout/student is a Hindu/Sikh and 2 if Muslim.
  • (vii) “Stage of Education last attended” (E_stage). This variable takes 1 if last class attended by dropout/student is in the primary stage (Class I-V) and 2 for higher stages of school.

    (viii) “Cared for doing well in studies” (Care_s)_: This variable takes 0 if the answer is “no” and 1 if “yes”.

  • (ix) “Standard of Living” (Std_liv): This variable takes 2 if the standard of living is “middle” and 1 if “low”.
  • (x) “Total Fee” (Totfee): It measures annual school expenditure
  • incurred by the student/dropout and is the sum of tuition fee and annual admission fees.

    (xi) Variable sex is coded as 1 for males and 2 for females.

    Logistic Modelling of Dropouts/Students

    Result of logistic regression is placed in Table 8. The Hosmer Lemeshow goodness of fit test provided a p-value of 0.992 indicating that the fit is good. Additionally, the classification table provided further evidence on adequacy of the model with correct prediction over 82 per cent (the cut off is taken at 0.78 following the procedure outlined by Crammer (1999) since the sample is an unbalanced one). The coefficient estimates, their standard errors, p-values for testing their significance (as judged by Wald’s statistic) and the change in odds corresponding to unit increase in the dependent variable is given in the Table 8.

    V Discussion

    A look at the column of odds ratio shows interest in studies (cared for doing well in studies) as the most important predictor associated with dropping out of school. Compared to those with an interest in studies, those who did not have interest were 7.7 times more likely to dropout. It is pertinent to note here that this is consistent with the individual perspective framework of Rumberger (2001). This framework views the attitudes and behaviours of students through the concept of student engagement. One dimension of engagement is reflected in student’s attitude and behaviours within the formal aspects of school (the other dimension being informal ones like peer and adult relationship). The variable “cared for studies” reflects an attitude towards education and as the odd ratio shows, this is one of the most important determinants of school dropouts.

    As a student graduates from primary school to a higher stage of school, the odds of dropping out increases 2.7 times. Of late there has been an overwhelming emphasis on primary education perhaps in line with the UN’s millennium development goals. Both the government and NGOs have been active in this area. Consequently opportunities for primary education have expanded at a pace faster than non-primary education. Borooah (2003) studied a large Indian database (constructed by NCEAR) and observed that while only 11 per cent of children lived in villages without a primary school, 30 per cent lived in villages with or without a middle school. A similar picture is reflected in urban areas. A neighbourhood primary school is frequent while the same cannot be said about upper primary schools. Parents have started recognising the value of primary education and irrespective of their economic status, are nonetheless eager to send their children to primary school [Ramchandran and Saihjee 2002]. The fact that the likelihood of students dropping out in upper primary stages is more than that in primary stages is not a very happy

    Table 7: Means and Standard Deviations of Explanatory Variables

    Variable Mean Std Dev Variable Mean Std Dev
    E_s t age S e x Religion TOT_SIB Father_e Mother_e C a s te 1.40 1.51 1.20 2.88 8.61 1.95 1.58 .49 .50 .41 1.62 2.90 3.36 .49 Std_liv TOTFEE (in Rs) Car e _s B_wat W_par TTO S TG 1.39 146.88 .3443 .73 .39 .60 .69 .49 274.27 .4754 .89 .73 .49 .74

    Economic and Political Weekly December 23, 2006

    situation as the benefits of primary schooling can get consolidated only if upper primary education is perssued. Five years of primary education is insufficient to ensure significant value addition and in many cases even inadequate to ensure retention of basic literacy and numerical skills acquired during primary schooling.

    The data set used in this analysis had respondents from three religions, viz, Hindus, Muslims and Sikhs. I have clubbed Hindus and Sikhs into a single group for the purpose of this analysis in view of present day realities. Results show that the odds of a Muslim student discontinuing school is 1.9 times that of Hindus/ Sikhs. The relationship is only weakly significant at 10 per cent. Investigations by Borooah (2003) has reported a significantly lower likelihood of continuing school in case of Muslims. Similar results have also been reported by Shariff (1995) and Sengupta and Guha (2002). The fact that dropout rate of Muslims is higher in India has also been borne out by the analysis carried out by Bhat and Zavier (2005) using the 1991 census data as well as NSSO 43rd (1987-88) and 50th (1993-94) rounds data. Why are their rates higher? Since higher educated parents tend to influence educational attainment positively, Bhat and Zavier have argued that communities that took to education earlier caused the advantage to be passed on to the next generation. Higher illiteracy or educational backwardness of Muslims is a legacy of the past. Consequently in urban India, following independence, Hindus were in a better position to take advantage of opportunities for secondary education than Muslims who lagged behind in primary education and literacy [Bhat and Zavier 2005].

    The total number of siblings has been found to be a highly significant predictor of dropping out. An increase in family size by one increases the odds of dropping out 1.7 times. There are essentially two diametrically opposing views on the association between family size and educational attainment. A number of empirical investigations have concluded that larger families present educational disadvantages to their children as compared to smaller families [Knodel and Wongsith 1991; Anh et al 1998]. These studies have confirmed the sibling rivalry and the dilution hypotheses. However, Chernichovsky’s (1985) study in Botswana has challenged this notion with two convincing arguments. First, he found that a larger number of school age children within a household enhances the likelihood of a child to be enrolled in school. This finding, as argued by Chernichovsky, reflects lower demand for labour of each individual child at home when more children are available, and reduces the indirect cost of educating a child. Often parents in developing countries assign different roles to their children. This has been called child specialisation. The phenomenon involves certain siblings going to school while others work. Many times this depends on the birth order where the oldest is the one who attends school. Second, the consideration of family type suggests that the extended family could mitigate against the family size effect. Therefore, by having grandparents residing in the same household, children are more likely to be enrolled in school than those in nuclear families. In the Indian context, perhaps the negative association holds as our analysis has shown. Borooah (2003) has obtained similar evidence. Krishnaji (2001) found children-women ratio to be an influencing factor in the enrolment and continuation of girls in school in his study in the state of Andhra Pradesh in India.

    Socio-economic Status (SES) most commonly measured by parental education and income is a powerful predictor of school continuance [Ekstrom et al 1986, Rumberger 1983]. Higher or more educated parents may simply serve as better role models influencing their children’s aspirations for more schooling [Rumberger 1983]. My analysis confirms that father’s level of education is significantly related to dropout behaviour. For each higher class of father’s education, the likelihood of a student dropping out reduces by 16 per cent. However I did not find mother’s education important. This is a little disquieting since quite a few studies have reported evidence which suggests mother’s education to be more important than father’s education with regard to retention in school [Psacharopoulos and Arrigada 1989, Sathar and Lloyd 1994]. Sengupta and Guha (2002) in their analysis of female dropouts in the state of West Bengal (of India) have reported something similar to my results. To quote these, “interestingly enough, mother’s primary education or middle level schooling did not have significant influence on dropouts in the city possibly due to strong, positive influence of urban residence on school participation”. With regard to father’s education, their finding is similar to mine. Father’s education level exerted a strong positive influence on retention in school. Perhaps more research is required to establish the association between mother’s

    Table 8: Results of Logistic Regression

    Independent variables Estimate of Coefficient Std Error p-value (Wald’s Statistic) Odds Ratio 95.0 Per Cent CI for Odds Ratio Lower Upper
    E_stage_Primary (Ref: Upper Primary) SEX_Males ( Ref: Females) Religion _Muslims (Ref: Hindus/Sikhs) TOT_SIB Father_e Mother_e Caste_backward (Ref: Non backward) Std_liv_Low (Ref: Medium) TOTFEE .980 .208 .665 .540 -.177 -.032 1.171 .702 .000 .299 .272 .376 .102 .050 .043 .283 .274 .001 .001 .445 .077 .000 .000 .459 .000 .010 .574 2.665 1.232 1.945 1.716 .838 .968 3.226 2.018 1.000 1.482 .722 .930 1.406 .760 .890 1.854 1.180 .999 4.793 2.101 4.067 2.094 .923 1.054 5.614 3.449 1.002
    Care_s_NO (Ref: YES) B_wat (rarely or never)R B_wat (does it sometimes) B_wat (does it almost regularly) W_par (rarely or never)R W_par (does it sometimes) W_par (does it almost regularly) TTO (regularly or sometimes)R TTO (rarely or never) STG (almost daily or regularly)R STG (rarely or never) STG (sometimes) Constant 2.042 .310 1.308 -.187 1.621 .629 1.388 1.301 -2.700 .267 .358 .359 .441 .571 .314 .382 .357 .650 .000 .386 .000 .672 .005 .045 .000 .000 .000 7.707 1.364 3.699 .829 5.059 1.875 4.005 3.672 .067 4.569 .677 1.830 .349 1.651 1.014 1.895 1.825 13.000 2.748 7.478 1.969 15.501 3.469 8.464 7.389
    Note: R stands for reference category.
    Economic and Political Weekly December 23, 2006 5261

    education and retention in the urban school set up of India before any firm conclusions can be drawn. Family income which I measured through its proxy “standard of living” was found to be highly significant in predicting school dropouts even after controlling for other factors (a proxy variable was used because family income is notorious for under reporting). A student in a family with standard of living “low” had twice the chance of dropping out as compared to a student in “middle” category.

    Caste is a historical legacy in India. After independence (in 1947), the government did undertake a major effort to try and improve the lot of backward castes. Special schemes and programmes targeted at these sections of the society still operate. Though there has been some improvement, the accumulated baggage of many generations will take longer to go away. Even though the position of the backward castes has improved, discrimination still continues. My analysis confirms this situation. Even after 55 years of development experience, caste remains a barrier to continuing school education. Compared to a child from, the privileged castes, a backward class student is 3.2 times more likely to dropout.

    The Constitution distributed the subjects of governance into three lists, the union list, the state list and the concurrent list. Education being in the concurrent list (since 1976), both centre and state governments has been spending a huge amount of money. Revised budget estimates (revenue expenditure) of union government for FY 2004-05 was 80 billion rupees for ‘Elementary Education and Literacy’ and 52.242 billion rupees for ‘Secondary and Higher Education’ (from in March 2005). Assam (the area of survey, Guwahati is its capital city) made a budget provision of 32.96 billion rupees on education for FY 2004-05. This is 30 per cent of its total budgeted revenue expenditure for the year (1 US $ = Rs. 45 ± 2). As I mentioned earlier, some 98 per cent of educational institutions in Assam are directly or indirectly supported with funds from state budget. During 1998-99, 96.44 per cent of primary schools of the state of Assam were government schools run and managed by the education department of the state government [Aggarwal 2000]. In these schools, education is virtually free. Tuition fees etc are nominal or almost non-existent. In many schools, some books are also provided free of cost or are highly subsidised by the government. Subjects of our sample were drawn entirely from G and GA schools so that both dropouts as well as students paid minimally for their school education. In this background, it is not surprising that the variable “totfee” was not found to be significantly associated with dropping out. What this result says is that cost of school education is not a factor significantly influencing dropout behaviour (perhaps this observation needs to be qualified with the words “in government and government aided schools” as subjects of my sample are entirely from such schools). The variable “totfee” includes annual admission fees and monthly fees.

    With a p value of 0.445, my analysis did not indicate sex to be significantly associated with dropping out. However the odds ratio indicates that boys have 20 per cent higher likelihood of dropping out as compared to girls. Since female dropout rates have consistently been higher in the Indian context (see Table 1), this finding appears to be a misfit. Borooah (2003) also reported “that the probability of being enrolled at school and then continuing in school was lower for girls than that of boys”. The apparent misfit can possibly be explained by the fact that dropout statistics for the state of Assam is somewhat mixed with higher dropout rates for males in the primary stages.

    My analysis supports the hypothesis of association between familial duties and dropping out. I included two variables, viz, “Bringing water for family use” (B_wat) and “Working with Parents in Outdoor Activities” (W_par) to examine this hypothesis. Both these variables are significant at 1 per cent thereby establishing strong association between familial duties and dropping out. Compared to the children who never do or rarely do such duties, those who do them regularly are 3.7 times and 5 times respectively more likely to dropout. Why are such duties necessary? As per census 2001, only 37.9 per cent of households of Assam had the source of drinking water within their premises. Perhaps more importantly, 22.4 per cent had the source of drinking water far away from the household premises. This indicates that for almost two-third of households of the state, bringing water is a major household chore. While households with a high standard of living can afford to purchase water (or perhaps install costly equipment to ensure availability of water within their premises), the rest are particularly disadvantaged. Family members are forced to share this job. In this part of the country, it is not an uncommon sight to see adults and young boys/girls carrying water in small containers. Additionally in Indian homes, working with parents or helping them in their outdoor activities is often considered an essential attribute of a well behaved ward. Parents endeavour to instil these values in their children. While the positive effect of children lending a helping hand to their parents can hardly be denied, the problem arises when the same is required to be done at the cost of school or study time. The burden of familial duties distracts and gradually weans away the child from educational activities leading to school discontinuance.

    A clarification on the direction of relationship is in order here. One could ask whether household chores caused school discontinuance or the other way round. The former is true. During data collection, respondents were specifically requested to provide us with a portrait of dropouts with information relating to their activities and conduct during the period prior to the time around which school discontinuance occurred. Considerable effort was expended to ensure that this aspect sank in before any response was elicited.

    The hypothesis of parental bonding has also been accepted. While the variable “Taking them out” (TTO) is significant at 5 per cent, the other variable “Story telling and gossiping” (STG) is weakly significant at 10 per cent. This implies that the depth of parental bonding does indeed influence school retention. Child psychologists these days caution that mere presence (of parents) at home is not enough. Spending quality time is the key. Though this message is prescribed in the general sense of child well being, results of my logistic regression confirms that the same holds good equally well even for school continuance. Parents need to involve themselves in the activities of their children and allocate some time for them exclusively. Through such activities, parents provide emotional support to their children. Bonds with parents get stronger in the process. I believe such bonding will also have other positive offshoots. First, spending more time with children would immediately mean parents would be in a better position to monitor activities of their children. Parental supervision improves. They are in a better position to ensure that creative energies of their children are properly channelised perhaps redirecting them into educational activities in the process. I have noted in Table 5

    Table 9: Dropout Rates for the State of Assam during 2002-03

    Primary (Class I-V) Class I-VIII Class I-X Boys Girls Total Boys Girls Total Boys Girls Total

    62.51 59.63 61.17 67.07 70.85 68.76 74.28 75.65 74.91

    Source: Rajya Sabha (upper house of Indian Parliament) Unstarred Q No 355 dated 16-12-04.

    Economic and Political Weekly December 23, 2006

    that in my sample, just 22.2 per cent of dropouts reported that parents (either or both) first proposed to them to discontinue school. Parents do desire that their wards continue in school and better parental bonding has the potential to ensure that their views find better acceptability among their wards. Second, improved bonding would mean closer interaction with their children. This means parents will be able to keep a tab on undesirable peer relationships which are often harmful. This is particularly important for families which reside in low quality neighbourhoods. The relationship between undesirable peer relationship, smoking, drugs, teenage pregnancy has already been well documented [Rumberger 2001].


    The analysis reveals that familial duties and parental bonding both play an important role in determining dropout behaviour. This influence exists even after controlling for a host of other well known socio economic and educational factors. Two policy implications immediately crop up. First, this result shows the importance of meeting basic necessities of citizens. Water is a fundamental human necessity. Access to clean and safe drinking water is not only a natural fundamental right of citizens but also a fundamental duty of all governments towards their citizens. This duty needs to be performed on a priority basis. The importance of pure drinking water vis-a-vis health hardly requires elucidation. This study illustrates one more reason why governments should endeavour to provide safe and plentiful drinking water within the household premises. Second, the publicity campaign by governments, NGOs with regard to importance of education perhaps requires a little fine tuning. I have illustrated how the regularity of working with parents in outdoor activities as well as parental bonding plays a significant role in dropout behaviour. These aspects needs to be built into publicity campaigns. Parents and guardians need to be told to find more time for their children. Not only do they need to give their children company, they should also interact with them intensely. Positive energies of future adults cannot only be canalised in the right direction in this endeavour but additionally, distracting and perhaps harmful influences on their children can also be kept at an arm’s length in the process.



    [An earlier version of this paper was presented at the ‘National Conferenceon Population and Development in North-east India’ held in February 2004at North Eastern Hill University, NEHU. The conference was sponsored byInternational Institute of Population Sciences (IIPS), Mumbai, India.Data for this paper has been taken from a study titled ‘Mathematical Modellingof School Dropouts and Data Analysis’ carried out during 2001-03 at thedepartment of statistics, Gauhati University under a three member departmentalteam consisting of Pranita Sarmah, Labananda Choudhury and the author.The study was funded by the ministry of statistics and programmeimplementation, government of India. The author wishes to place on recordhis sincere appreciation and gratitude for the contribution made by the othertwo members of the study team in data collection and data management.]


    Anh, Truong Si, John Knodel and David Lam (1998): ‘Family Size andChildren’s Education in Vietnam’, Demography, 35, pp 57-70.

    Aggarwal, Yash (2000): How Many Pupils Primary Education in Five Years,

    National Institute of Educational Planning and Administration (NIEPA),

    New Delhi. Bhat, P N Mari and A J Francis Zavier (2005): ‘Role of Religion in Fertility

    Decline: The Case of Indian Muslims’, Economic and Political Weekly,

    XL, 5, pp 385-402.Borooah, Vani K (2003): ‘Births Infants and Education: An Econometric

    Portrait of Women and Children in India’, Development and Change,

    34, pp 67-102.

    Chernichovsky, D (1985): ‘Socio-economic and Demographic Aspects ofSchool Enrolment and Attendance in Rural Botswana’, Economic Development and Cultural Change, 32, pp 319-32.

    Colcough, C (1982): ‘The Impact of Primary School on Economic Development:A Review of Evidence’, World Development, 10(3), pp 167-85.

    Crammer, J S (1999): ‘Predictive Performance of the Binary Logit Modelin Unbalanced Samples’, The Statistician, 48, pp 85-94.

    Dornbusch, S M, P L Ritter, P H Leiderman, D E Roberts and M J Fraleigh(1987): ‘The Relation of Parenting Style to Adolescent SchoolPerformance’, Child Development, Vol 58, No 5, pp 1244-57.

    Ekstrom, R B, M E Goertz, J M Pollack and D A Rock (1986): ‘Who DropsOut of High School and Why? Findings from a National Study’, Teachers College Record, 87, pp 356-73.

    Finney, P (1993): ‘The PTA/Newsweek National Education Survey’,Newsweek, May 17.

    Henderson, A T and N Berla (1994): ‘A New Generation of Evidence: TheFamily Is Critical to Student Achievement’, St Louis, MO: DanforthFoundation and Flint, MI: Mott (CS) Foundation.

    Jimerson, S, B Egeland, A Sroufe and B Carlson (2002): ‘A ProspectiveLongitudinal Study of High School Dropouts Examining Multiple Predictorsacross Development’, Journal of School Psychology, 38(6), pp 525-49.

    Keng, Chansopheak (2004): ‘Household Determinants of SchoolingProgression among Rural Children in Cambodia’, International Education Journal, 5(4), pp 552-61.

    Kingdon, G G (1999): ‘Gender Gap in Indian Schools: Is Labour Marketa Factor’? Id21 Insights, Issue 29.

    Knodel, John and M Wongsith (1991): ‘Family Size and Children’s Educationin Thailand: Evidence from a National Sample’, Demography, 28, pp 119-31.

    Krishnaji, N (2001): ‘Poverty, Gender and Schooling: A Study to TwoDistricts in Andhra Pradesh’ in A Vaidyanathan and P R Gopinathan Nair(eds), Elementary Education in Rural India: A Grass Toots View, SagePublication, New Delhi.

    Louis Harris and Associates (1993): Metropolitan Life Survey of the AmericanTeacher 1993: Violence in American Public Schools, Laura J Colker, New York.

    NSSO (National Sample Survey Organisation) (1997): Report No 412,‘Economic Activities and School Attendance by Children of India’, May,Department of Statistics, Government of India.

    Psacharopoulos,George and A M Arriagada (1989): ‘The Determinants ofEarly Age Human Capital Formation: Evidence from Brazil’, Economic Development and Cultural Change, 37(4), pp 683-708.

    Ramchandran, V and A Saihjee (2002): ‘The New Segregation: Reflectionson Gender and Equity in Primary Eduation’, Economic and Political Weekly, 37(17), April 27, pp 1600-13.

    Rumberger, R (1983): ‘Dropping out of High School: The Influence of Race,Sex and Family Background’, American Educational Research Journal, 20, 2, pp 199-220.

    – (2001): ‘Why Students Dropout of School and What Can Be Done’, paperpresented at the conference ‘Dropouts in America: How Severe Is theProblem? What Do We Know about Intervention and Prevention’, Harvard University.

    Sarin, Ankur (2003): ‘Does Economic Inequality Matter Even When AverageIncomes Are Low: An Analysis of Educational Attainment in Rural India’,paper presented at the Population Association of America Annual Meeting,2003.

    Sathar, Zeba A and Cynthia B Llyod (1994): ‘Who Gets Primary Schoolingin Pakistan: Inequalities within Families’, The Pakistan DevelopmentReview, 33(2), pp 103-34.

    Schneider, B H (1993): Children’s Social Competence in Context, PergamonPress, Oxford.

    Sengupta, P and J Guha (2002): ‘Enrolment Dropout and Grade Completionof Girl Children in West Bengal’, Economic and Political Weekly, 37(17),April 27, pp 1621-37.

    Shariff, Abusaleh (1995): ‘Socio-Economic and Demographic Differentialsbetween Hindus and Muslims in India’, Economic and Political Weekly,18, November 18, pp 2947-53.

    Steinberg, L, S D Lamborn, N Darling, N S Mounts, S M Dornbusch (1994):‘Over – Time Changes in Adjustment and Competence among Adolescentsfrom Authoritative, Authoritarian, Indulgent and Neglectful Families’,Child Development, Vol 65, No 3, pp 754-770.

    Vaid, D (2004): ‘Gendered Inequality in Educational Transitions’, Economic and Political Weekly, 34, No 35, pp 3927-38.

    Weber, J M (1989): ‘Identifying Potential Dropouts: A Compilation andEvaluation of Selected Procedures’, Centre on Education and Trainingfor Employment, Columbus, Ohio.

    World Bank (1997): Primary Education in India, World Bank, Washington DC.

    Zsolnai, Aniko (2002): ‘Relationship between Children’s Social Competence,Learning Motivation and School Motivation’, Educational Psychology,Vol 22, No 3, pp 317-29.

    Economic and Political Weekly December 23, 2006

    Dear reader,

    To continue reading, become a subscriber.

    Explore our attractive subscription offers.

    Click here


    (-) Hide

    EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

    Back to Top