Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Results of a school lunch programme in India

Sneh Rewal
CARE-lndia, Greater Kailash, New Delhi, India


Some doubts have been raised about the effectiveness of feeding programmes for school-age children. Children at this age have survived the malnutrition and disease of the younger years of life. At school age, they are relatively less vulnerable to malnutrition than are pre-school children and are, therefore, not considered a high-risk group. Critics of school lunch programmes have drawn attention to their cost, to the limited availability of resources, and to inconclusive evidence of impact. It is not easy to deny such criticism in the absence of sufficient data. Is lack of information in itself, however, enough reason to conclude that nutrition intervention will not benefit this group? Equally, there is insufficient data to indicate that such programmes do not play an important role in correction of the nutritional inadequacies in the earlier years of life that result in poor growth, erratic school attendance, and reduced intellectual performance.

A closer review of the available information on the effectiveness of the school lunch programme is desperately needed before any useful conclusions can be drawn. This paper attempts to use recently collected data to derive analytical results and question the validity of certain widely held views about the utility and effectiveness of the programme.

The information examined in this paper pertains to children participating in the Mid-day Meals (MDM) Programme in Madhya Pradesh State in central India. The programme has been in operation for the past 15 years and covers all the schools under the Department of Tribal and Harijan Welfare in 21 districts. The premise underlying the programme was that a meal distributed free at the school, eaten in addition to home food, would result in an improvement in nutritional status and would act as an incentive for increasing contact with the school system. This group of children is easily accessible through the operational mechanism of the school system. Under this school lunch programme, a beneficiary child receives a meal prepared from 80 grams of grain and 7 grams of oil, providing 312 calories and 14 grams of protein (1). The annual target number of feeding days is 180.


The evaluation study covered 3,839 children randomly selected in 396 schools in all 21 programme districts. Dietary data were collected on a subsample of 2,045 children. The analysis was done within the group of programme participants. Due to the absence of baseline and longitudinal data and also to the lack of a matching control group, the children from high-programme-efficiency schools with a greater exposure to the programme were compared with children from low-efficiency schools with less exposure to the programme. Because of the unavailability of prerequisite information on programme efficiency, the schools were grouped de facto at the time of analysis. The cut-off points of the four programme-efficiency groups were decided arbitrarily as: low (0 to 60 per cent), medium 161 to 85 per cent), high (86 to 95 per cent), and very high (96 per cent and above), so as to have an almost equal number of schools in each of the groups. Programme efficiency was computed as the percentage of actual feeding days against the total number of school days.


The food given at the school was the first meal of the day for 20 per cent of the children; they had come to school with an empty stomach. In a majority of the schools, the meal was distributed at the lunch-time break between 1.00 2.30 p.m. Most of the children (79 per cent) ate food on the school premises. The same percentage of children indicated that they themselves ate all the food they were given. The majority of the children who did not eat the entire quantity of food at school shared it with their family members, and especially with younger siblings. The school meal provided 359 calories and 19 grams of protein compared to the prescribed ration of 312 calories and 14 grams of protein. As the quantity of food distributed to all the children was the same irrespective of their age, the percentage of the recommended allowances of calories and protein provided by this meal was greater for the younger children.

The children were clearly better fed on the days they ate the school meal. When the children did not get a school meal, they ate "some" food at home instead; the nutritional value of this additional home food was much less than that of the school meal. About one-third of the children missed a meal at home on the day they received food at school; these children ate more meals at home on a holiday. On average, 312 calories 187 per cent) of the total for the school meal supplemented calories obtained from food eaten at home, and 47 calories (13 per cent) replaced the home diet calories (table 1). On the other hand, all the protein in the school meal supplemented the home-food protein.

TABLE 1. Estimate of the Extent to Which Calories and Protein in School Meals Substitute for and Supplement Those from Home Food

  Source of Calories Utilization of School-Meal Calories
Home food School meal As substitute As supplement
kcal (% RDA) kcal (% RDA) kcal (% RDA) kcal (% RDA)
Children eating only home food 1,390 (71) - - - - - -
Children eating home food plus school meal 1,343 (68) 359 (18) 47 (2) 312 (16)
  Source of Calories Utilization of School-Meal Calories
Home food School meal As substitute As supplement
kcal (% RDA) kcal (% RDA) kcal (% RDA) kcal (% RDA)
Children eating only home food 38 (104) - - - - - -
Children eating home food plus school meal 38 (104) 20 (52) 0 (0) 20 (52)

The school meal minimized the difference between recommended standards and home food in calories as well as in proteins (fig. 1). It reduced the percentage of children receiving less than 75 per cent of the recommended daily allowance (RDA) of calories from 62 to 39 per cent and increased the percentage of children meeting the RDA requirements from 13 to 30 per cent. Similarly, the percentage of children receiving less than 75 per cent of the protein RDA was reduced from 30 to 4 per cent, and the percentage of children meeting the recommended requirement increased from 45 to 86 per cent.


The number of feeding days in a school and physical measurements of the children showed a high correlation. With an increase in the number of feeding days, nutritional status improved significantly (table 2). The children were found to be heavier and taller in the high-programme-efficiency schools in comparison with the children of the same age in the low-efficiency schools. The mean body weight and height expressed as a percentage of the reference standard was greater in the high-programme-efficiency schools. Analysis of data using three indices -weight for age, height for age, and weight for height- shows that a larger number of children (6 to 7 per cent more) fall into the category of "normal" in high-efficiency schools compared to low-programme-efficiency schools (table 3). Additionally the higher-efficiency schools had 2 to 3 per cent fewer children with third-degree malnutrition.

The persuasive evidence of a link between the number of feeding days and improvement in body weight was provided by classifying children by a combined weight-for-height and height-for-age index (table 4). The schools with the greater number of feeding days had a significantly higher number of children with "normal" and "marginal" nutritional status. There were considerably larger numbers of children with "wasted" and "wasted and stunted" growth in low-programme-efficiency schools. However, no statistically significant difference was observed in the number of children classified as "stunted"-with considerable height deficits and "normal" weight-between high- and low programme-efficiency schools.

FIG. 1. Effect of School Meal on Difference between Actual Daily Intake of Calories and Protein and Recommended Daily Allowance (RDA)

TABLE 2. Comparison of Nutritional Status of Children in Low- and High-Programme-Efficiency Schools

Nutritional-Status Indices Programme Efficiency
Very high,
> 96%
t Value between
Low and Very High
Weight as % of reference weight for age 68.4 68.7 69.7 70.9 4.20*.
Height as % of reference height for age 91.4 91.3 91.5 92.2 2.78**.
Weight as % of reference weight for height 84.8 85.5 85.9 86.9 3.33

N = number of children.
* Significant at 1 per cent level.
** Significant at 5 per cent level.

TABLE 3. Classification of Nutritional Status of Children by Programme Efficiency

Nutritional Status (% of standard Indices) Programme Efficiency
Very high,
> 96%
Weight for age        
< 60 22.1 23.0 24.6 19.9
61-69 37.0 33.9 31.0 28.5
70-79 26.0 27.4 23.9 31.1
80-89 10.6 11.0 14.0 12.0
> 90 4.4 4.7 6.4 8.6
Height for age        
79 1.6 1.5 1.9 1.9
80-84 8.0 9.4 11.1 9.1
85-89 27.0 27.7 24.5 22.1
90-95 37.5 36.0 33.9 33.5
> 96 25.8 25.5 28.6 33.5
Weight for height        
<69 1.5 1.0 1.1 1.3
70-79 22.1 20.6 16.9 15.1
80-84 25.2 26.6 24.9 26.7
85-89 27.8 26.8 28.3 28.1
> 90 23.4 25.1 28.1 28.8

N = number of children.
Values given are percentages of children in each category.

TABLE 4. Classification by Combined Weight-for-Height and Height-for-Age Index-by Programme Efficiency

Combined Indices Programme Efficiency
(N = 928)
(N = 1,226)
(N = 854)
Very high,
> 96%
(N = 794)
Normal and marginal (weight for
height > 80%; height for age > 90%)
47.0* 47.8 51.8 55.0*
Stunted (weight for
height > 80%; height for age <90%)
29.3 30.8 30.7 28.6
Wasted (weight for
height < 80%; height for age > 90%)
16.3* 13.6 10.9 11.8*
Wasted and stunted (weight for
height < 80%; height for age < 90%)
7.3 7.8 6.9 4.5*

N = number of children.
Values given are percentages of children in each category * Differences are significant at the 0.05 per cent level

TABLE 5. Effects of Selected Variables on Nutritional Status

  Age School
Sex Type of family Education of mother Education of fother Size of landholding Home caloric intake Programme efficiency Months of partici- pation
A. Standardized Regression Coefficient
Weight for 0.155 (0.018) (-0.005) 0.047 ( - 0.010) ( 0.001) 0.050 ( 0.017) - - 0.032
height 0.137 (-0.028) ( 0.008) 0.043 ( - 0.006) (0.004) 0.049 (- 0.016) 0 075 0.082 0.043
Weight for 0.445 0.324 - 0.057 (0.001 ) ( 0.029) - 0.055 (0.024) 0.081 - - 0.102
age 0.460 0.300 - 0.054 (-0.004) ( - 0.026) 0.050 (0.022) 0.082 0.072 (0.046) 0.109
Height for -0.508 0.350 - 0.072 (-0.022) ( - 0.024) 0.056 (0.004) 0.103 - - 0.133
age -0.510 0.349 0.070 (-0.024) ( - 0.023) - 0.054 (0.002) 0.103 (0.032) (0.005) 0.135
B. Regression Coefficient
Weight for 0.046 (0.096) ( 0.077) 0.712 ( 0.195) ( 0.012) 0.172 (- 0.000) - - 79.122
height 0.041 (0152) (-0.013) 0.647 ( 0.116) (0.046) 0.168 (- 0.000) 0.082 0.0294 77.721
Weight for -0.212 2.782 - 1.534 (0.026) ( 0.862) 0.900 (0132) 0.002 - - 84.290
age -0.217 2.576 - 1.453 (-0.090) ( - 0.782) - 0.819 (0.122) 0.001 0.035 (0.026) 81.994
Height for 0.121 1.516 - 0.963 (-0.270) ( - 0.363) - 0.464 (0.010) 0.001 - - 100.588
age -0.121 1.511 - 0.948 (-0.301) ( - 0.355) - 0.447 (0.007) 0.001 (0.008) (0.001) 100.03


The programme experience indicates, in brief, that regularly distributed food eaten by school-age children does have a positive impact on body weight.

The regression analysis also supports the above evidence. In the three equations, per cent weight for height, per cent weight for age, and per cent height for age were hypothesized to be a function of age, sex, grade in school, type of family, education of parents, size of landholding, home calorie intake, months of programme participation, and programme efficiency (table 5). In this analysis, months of participation in the programme and programme efficiency were found to be significant predictors of weight for height. Per cent weight for age was not influenced by months of participation; it was explained by programme efficiency, socio-economic, demographic, and dietary intake variables. Height for age was totally independent of the two programme participation variables, This finding was not surprising, as weight is an indicator of the current nutritional status and height conveys information primarily of past nutritional status. With an average of 31 months' programme participation, body weight has a better probability of being significantly improved. Also, it may be observed that in the first two equations {weight for height and weight for age) the effect of socio-economic variables is reduced with the inclusion of programme variables.


Apparently, the most appealing aspect of the MDM programme is its ability to satisfy hunger for the children who come to school with an empty stomach or after only a little breakfast. Not only does short-term hunger affect school performance (2), but it also has some adverse effects on emotional behaviour (3; 4).

From the data available, it is evident that the food reached the intended beneficiaries. This level of achievement has not yet been attained in pre-school feeding programmes (5). Moreover, there is only a little leakage in the MDM programme, and even the leakage has an important benefit, as it is mostly used for the younger pre-school age siblings. What is more, almost the entire quantity of the school meal is used as a supplement to home food. The gap between the total home calorie intake and the RDA is reduced to from 10 to 20 per cent.

This supplementary food eaten on a regular basis had a positive impact on nutritional status; the gain was manifested first in body weight and later in height increments. The difference in the nutritional status of the two groups of children was statistically significant but not dramatic. If the participant children with greater exposure to the programme had been compared to children with no intervention rather than with children with less exposure to the programme, the difference could have been striking (6).

The question often raised is whether this difference can affect school attendance and school performance. The available data did not suggest any clear conclusions on school attendance. The most rigorous test of the programme's impact on attendance measured by number of days present in the school during a six-month period did not indicate any trend. A closer look at the data indicates that comparison of school attendance in the arbitarily grouped schools is invalid. The low-programme-efficiency schools are not characteristic of a control group. These schools were certainly distributing food for a comparatively fewer number of school days, but were in effect close to meeting the target of feeding days set for the school year. Also, the feeding programme operating for the past 15 years in Madhya Pradesh State has become a part of the school system, and a few days' interruption or irregular distribution is not likely to affect school attendance. However, other investigations detected a significantly positive impact. In a study conducted by CARE in Karnataka, a south Indian state, it was found that month by-month variations in enrolment levels and attendance rates tended to be significantly less in school with a mid-day meal than in schools not providing meals (7).

Regarding school performance, no information was available from the study in Madhya Pradesh because school-type tests or examinations are not given. However, studies conducted in developed countries are conclusive about the need to make up for nutritional deficiencies in order to improve the quality of educational results (8; 9).


The data that are available do provide sufficient evidence to indicate that the programme can make a difference. Much-needed dietary supplements can be delivered to children through such programmes. Food reaching the intended beneficiaries is consumed by and large by them, and thus the school meal as a supplement to home food has the desired effect. This nutrition intervention can make up for the deprivation, neglect, and nutritional impoverishment of the earlier years of life. If pre-school child nutrition interventions are useful in bringing children through the most critical period, the MDM programme sustains and develops their nutritional and health status at school age. As our interest is not only in the number of surviving children but in a healthier child population, the MDM programme has an important role in strategies concerned with malnutrition, ill health, and poor intellectual performance.


1. R.P. Devadas and A. Radharokmani, The School Lunch Program
(Ministry of Education, New Delhi, India, 1966), p. 30.

2. N.E. Matheson, "Mid-Morning Nutrition and Its Effects on
School-type Tasks" [University of California, 1970, unpublished).

3. M, Kiester, "Relation of Mid-morning Feeding to Behavior of
Nursery School Children," J. Amer. Dietet. Assoc., 26:25 (1950).

4. D,A, Laird, M. Levitan, and V.A. Wilson, "Nervousness in
School Children as Related to Hunger and Diet," Med. J. Record
134:494 (1931).

5. T. Gopaldas, N. Srinivasan, I. Varadarajan, A.G. Shingweka, R.
Seth, R.S. Mathur, and V. Bhargava, Project Poshak, vols. I and II
(CARE-lndia, New Delhi, 1975).

6. D.T. Campbell and J.G. Stanley, Experimental and Quasi-Experimental
Designs for Research (Rand McNally College Publishing Company,
Chicago, 1960).

7. "School Feeding in Karnataka: Impact on Enrollment and
Attendance" (CARE-India, New Delhi, 1977),

8. F. Leninger, "Relation of the Use of Milk to the Physical and
Scholastic Performance of Undernourished School Children,"
Amer. J. Public Hlth., 25:555 (1933).

9. E. Pollitt, D. Greenfield, and R. Leibel, "Effects of Short-Term
Fasting on Cognitive Test Performance among 9 to 11-Year-Old Children,"
paper presented at the meeting of the American Academy of
Child Psychiatry, San Diego, Calif, USA, Oct. 1978.

Contents - Previous - Next