Contents - Previous - Next

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


The initial observation sample was 120 children. Seven children were lost from the study, and the resulting sample consisted of 57 undernourished and 56 adequately nourished children.

The children's anthropometry and social background are shown in table 4. As would be expected from the selection criteria, the undernourished children were significantly smaller in all anthropometric measures than the adequately nourished children. The mean age of the undernourished children was six months greater than that of the adequately nourished children (p < .01) and they had fewer schoolbooks (p < .05). The other two socio-economic ratings were not significantly different in undernourished and adequately nourished children.

TABLE 4. Anthropometric and socio-economic characteristics of children according to nutritional status

Characteristic Nutritional status
(29 M, 28 F)
Adequately nourished
(28 M, 28 F)
Mean SD Mean SD
Age (yr) 9.68 0 42 9.18 0.77**
Height-for-age (z-score) -1.30 0.54 0.58 0.91***
Weight-for-age (z-score) -1.56 0.28 0.30 0.56* * *
BMI (kg/m˛) 14.43 0.92 16.45 1.50**
Book score 4.61 3.75 6.28 3 87*
Housing score (0-8) 2.67 1.86 3.00 1.89
Uniform score (0--4) 2.93 1.08 3.23 0.97

*p < .05, **p < .01, ***p < .001 by Student's retest.


The correlations among the behaviours summed over four days were examined for the teaching and set task situations separately. In the teaching situation, children who were more on task participated more (r= .72, p < .001), and spoke (r= -.48, p < .001) and moved (r= -.21, p < .05) less, than children who were less on task. Children who talked more also moved more (r = .25, p < .01) and participated less (r= -.20, p < .05). During the set task situation, the children who were less on task talked more (r = -.65, p < .001) and moved more (r = -.28, p < .01). Also, as they talked more, they moved around more (r = .40, p < .001).

Correlations were calculated between the behaviours summed over the four days and the anthropometric measures and age. In the teaching situation there were no significant associations between the behaviours and the anthropometric measures and age. In the set task situation there was a significant positive correlation between attention to task and the children's height-for-age (p < .05) and BMI (p < .05) (table 5). There were significant negative associations between all three anthropometric measures and gross motor behaviour in the set task situation (p < .05). Thus, children with poorer nutritional status tended to move more at their desks and in the classroom when left to work on their own. When the adequately nourished and undernourished groups were considered separately, the same pattern of correlations was observed in the adequately nourished children but not in the undernourished children. It is possible that the restricted range of nutritional status accounted for the latter finding.

TABLE 5. Pearson product-moment correlations (r) of behaviours over four days in the set task situations with anthropometry and age on enrollment

Behaviour Height-for-age Weight-for-age BMI
ON TASK .16* .14 .16*
TALKS -.07 -.05 -.12
GROSS MOTOR -.18* -.21* -.17*

*p < .05.

The effect of nutritional status was further examined in a repeated-measures analysis of variance, with the behaviours with and without breakfast as the within-subject factor and nutritional group as the between-subject factor. There were no main effects of nutritional group on any of the behaviours in the teaching situations. However, in the set task situation there was a nutritional group effect on gross motor behaviour (F = 4.31, p < .05), with the undernourished group moving more often.

The frequencies with which the behaviours occurred in the teaching and set task situations are shown in tables 6 and 7. Repeated-measures analyses of variance for each behaviour, using behaviour with and without breakfast as the within-subject factor and school, sex, and nutritional group as the between-subject factors, were calculated for the two types of situations separately. In the set task situation, girls talked more (F = 6.90, p < .05) and had more gross motor activity (F = 6.23, p < .05) than boys. There were no significant interactions between sex and treatment. There were also no significant main effects of nutritional group or treatment on any of the classroom behaviours and no significant interactions between nutritional group and treatment in these analyses.

TABLE 6. Unadjusted medians and ranges of frequency of behaviours with and without breakfast in the teaching situation by school (undernourished and adequately nourished groups combined)

Behaviour by school Breakfast No breakfasta
School SV (n = 24)  
ON TASK 151.3 (18.0)b 137.3 (28.0)b
TALKS 4 (0-31) 6 (1-32)
GROSS MOTOR 8 (0-36) 14 (1-60)
PARTICIPATE/RESPONSE 45 (15-109) 49 (6-87)
School Ml (n = 22)  
ON TASK 123.2 (31.5)b 128.5 (32.8)b
TALKS 13 (1-38) 7 (1-22)
GROSS MOTOR 7 (1-17) 8 (2-16)
PARTICIPATE/RESPONSE 37 (2 86) 34 (6-139)
School GH (n = 18)  
ON TASK 125.9 (39.6)b 137.4 (34.2)b
TALKS 15 (5 41) 12 (0-34)
GROSS MOTOR 13 (3-40) 12 (2-42)
PARTICIPATE/RESPONSE 47 (16-83) 43 (13-101)
School AR (n = 49)  
ON TASK 129.3 (27.0)b 131.1 (28.2)b
TALKS 12 (0-59) 12 (0-74)
GROSS MOTOR 9 (1-34) 8 (1-20)
PARTICIPATE/RESPONSE 37 (8-143) 31 (6-112)

a. Children given one-quarter of an orange.
b. Mean (SD) number of 10-second periods "on tasks".

There were however, significant school effects and school-treatment interactions in both situations, indicating that the treatment affected children differently in the different schools. In the teaching situation there was a significant interaction effect in "on task" (F = 4.43, p < .01), "talks" (F = 3.33, p < .05) and "gross motor" behaviours (F = 2.71, p < .05). In the set task situation, there was a significant school-treatment interaction in the "on task" behaviour (F = 5.71, p < .001).

TABLE 7. Unadjusted medians and ranges of frequency of behaviours with and without breakfast in the set task situation by school (undernourished and adequately nourished groups combined)

Behaviour by school Breakfast No breakfasta
School SV (n = 24)  
ON TASK 143.1 (20.2)b 134.6 (30.9)b
TALKS 16 (1-71) 15 (4-71)
GROSS MOTOR 11 (1-33) 10 (3-50)
School Ml (n = 22)  
ON TASK 121.1 (29.5)b 136.4 (23.4)b
TALKS 25 (1-66) 19 (7-54)
GROSS MOTOR 6 (0-15) 6 (0-17)
School GH (n = 18)  
ON TASK 119.4 (22.8)b 135.4 (16.5)b
TALKS 33 (10-86) 24 (4-65)
GROSS MOTOR 9 (2-25) 7 (1-28)
School AR (n = 49)  
ON TASK 118.0 (22.7)b 112.9 (27.6)b
TALKS 34 (3-99) 32 (3-103)
GROSS MOTOR 9 (0-25) 10 (1-47)

a. Children given one-quarter of an orange.
b. Mean (SD) number of 10-second periods "on task."

Post-analysis of variance comparisons showed that during the teaching situation, the children's attention to task in school SV increased significantly with breakfast (t = 2.89, p < .01; see FIG. 1. Mean number of "on task" periods by school, with and without breakfast in the teaching situation), whereas the children's attention in the other schools did not change significantly. Similarly, in school SV, after the children had breakfast they moved less (t= 2.26, p < .05), whereas gross motor behaviour did not change in the other schools (see FIG. 2. Mean gross motor movements by school, with and without breakfast in the teaching situation). In contrast, in school MI, after having breakfast the children talked more to their classmates in the teaching situation (t = 2.20, p < .05), and in the set task situation children were less on task in schools MI (t = -2.73, p < .01) and GH (t = -2.58, p < .02). There were no significant changes in behaviour with breakfast in school AR.


To our knowledge this is the first study of the effects of breakfast on children's classroom behaviour in a developing country. The design of the study was particularly rigorous, with each child being compared with himself or herself with or without breakfast. In this way differences among classes, such as the teacher and peer group, were controlled. In addition the order of treatment was assigned randomly. We estimated that a difference of approximately 0.5 SD may have functional implications. The overall power to detect a treatment effect of this size on the "on task" behaviour at the .05 level was greater than 90% [34].

The reliability of the behaviours over two days was good, except for gross motor behaviour, which had only moderate reliability. This indicates that the time-sampling instrument was a reasonably reliable measure of the children's behaviour. The correlations among the behaviours were in the expected directions, with children who spent more time "on task" talking and moving less and responding to the teacher more. Children who talked more also moved more and responded to the teacher less. These associations give support to the construct validity of the behaviour measures. These correlations also indicate a pattern of behaviour in the classroom that may reinforce attention to task. Paying attention entails not only keeping the eyes focused on the task. but also more involvement with the teacher, less socialization, and sitting relatively quiet. A combination of these behaviours may thus contribute to better learning in class.


Undernutrition and behaviour

Over the four days of observation, shorter and thinner children were less attentive in class and, when left on their own to complete a task, were more likely to be restless and move around. Severely malnourished children have been reported to be more restless and to have poorer attention than matched comparison children [19, 20]. They have also been described as having attention deficit disorder [20]. Investigators in both Kenya [21] and Egypt [22] failed to find an association between anthropometry and children's classroom behaviour, although they found associations with dietary intake.

Although there were associations between behaviour and anthropometry in this study, the effects of breakfast were not different in undernourished and adequately nourished children. It is possible that an effect of nutrition might be found in populations with more severe undernutrition.


Treatment effects

The treatment effect of breakfast varied by school, which makes interpretation of the findings difficult. Breakfast improved the behaviour of children in school SV. After receiving breakfast the children were more attentive and less restless. Both of these behavioural changes would be considered favorable and should be conducive to improved learning [14].

In two of the other schools the children's behaviours actually deteriorated after they had breakfast. They paid less attention during the set task situation, and in one school they talked more to their classmates in the teaching situation after having breakfast. These behavioural changes are probably undesirable and unlikely to improve learning. Children in the remaining school were not affected when they had breakfast.

Close inspection of the schools showed that the structure of the school in which the children showed improvement with breakfast, school SV, was much better than that of the other three schools. School SV was the most recently built of the schools. Each class had its own room, and all of the children had their own desks and chairs (fig. 3). The desks were placed in well-spaced rows, and the classrooms were well lit and airy. In addition, the classroom atmosphere was reasonably quiet. In two of the other schools there were multiple classes in a room, separated at most by blackboards (fig. 4). The benches, which were designed for two children, were usually used by three or four. The noise level was high in these schools and there was an atmosphere of confusion. The fourth school, Ml, was the most remote and the smallest school. Although each class had its own room, there was a shortage of furniture. The limited seating was therefore placed close together, crowded into a small area of the room to accommodate as many children as possible. The rooms were also poorly lit and ventilated.

School facilities, such as furniture and equipment, have been found to be an important factor in predicting children's school achievement [1, 33]. In addition the physical arrangement of the classroom can affect children's classroom behaviour [34, 35]. We therefore hypothesize that the infrastructure of the school modified the effects of breakfast on the children's behaviour. It is possible that where classroom conditions were overcrowded and poorly controlled, the extra energy provided by the breakfast could not be channeled into appropriate classroom behaviour. Conversely, when classroom conditions were conducive to learning, improving the state of the children improved their behaviour.

We previously reported that the undernourished children in the present study showed an improvement in verbal fluency with breakfast [24]. It is likely that other, unmeasured cognitive functions also improved. It is therefore reasonable to assume that the state of the children improved with breakfast. However, this improvement was not consistently translated into improved classroom behaviour.

In conclusion, the provision of school breakfast improved the desired classroom behaviour of the children in the one school that was adequately equipped and organized. The increase in attention to task and reduction in restless behaviour should result in better school achievement [15], especially when accompanied by improved cognitive function. These improvements were not seen in the other schools, and behaviour actually deteriorated in two schools. Thus the benefit of school breakfast to children's behaviour may depend on the presence of satisfactory classroom infrastructure. There were only four schools in this study, and the findings need to be replicated before extrapolating them to other schools and countries. The implications of the findings are that where the infrastructure of the school is poor, it is necessary to improve the basic classroom amenities as well as providing school meals, if an improvement in school achievement is to be attained.


This study was supported by the Nestlé Nutrition Research Grant Programme and the Commonwealth Caribbean Medical Research Council. We would like to thank the observers, Sheron Barnes-Wilmot, Yasmin Williams, Althea Whitter, and Mala Siew-Narine. We would also like to thank the principals and children of the four schools for their cooperation.

Appendix: Definitions of behaviours

ON TASK (during teaching task situation): Child's eyes are focused on the teacher or the blackboard or the child is answering questions for 5 seconds or more of the 10-second block.

ON TASK (during set task situation): Child's eyes are focused on the blackboard (if lesson is on it) and child is writing, marking, or rubbing (erasing) for 5 seconds or more of the 10-second block.

TALKS: Child verbalizes or mouths words towards another child. Only initiations are recorded.

GROSS MOTOR: any marked motor movements from the child's starting position. Includes sitting to stand, sitting to turning 90 degrees around from waist, sitting to bending over from waist, sitting to stooping, and standing to walk two steps or more.

PARTICIPATE/RESPONSE: Any vocalizations by child towards the teacher, with face oriented towards the teacher. Also if child raises hand or recites, sings, or reads aloud when requested by teacher. Only initiations are recorded.


  1. Fuller B. Raising school quality in developing countries: what investments boost learning. Education and Training Department Discussion Paper EDT7. Washington, DC: Population and Human Resources Department, World Bank, 1985.
  2. Pollitt E. Malnutrition and infection in the classroom. Paris: UNESCO, 1990.
  3. Nokes C, Sandy DAP. Does helminth infection affect mental processing and educational achievement? Parasitol Today 1994;10(1):4-8.
  4. Simeon DT, Grantham-McGregor SM. Nutritional deficiencies and children's behaviour and mental development. Nutr Res Rev 1990; 3:1-24.
  5. Levinger B. School feeding programs in developing countries; an analysis of actual and potential impact. Aid Evaluation Special Study No. 30. Washington, DC: US Agency for International Development, 1986.
  6. Powell CA, Grantham-McGregor 8, Elston M. An evaluation of giving the Jamaican government school meal to a class of children Hum Nutr Clin Nutr 1983;37C: 381 -8.
  7. Meyers A, Sampson A, Weitzman M, Rogers B. Kayne H. School breakfast program and school performance. Am J Dis Child 1989;143:1234-9.
  8. Agarwal DK, Agarwal KN, Upadhyay SK. Effect of mid-day meal programme on physical growth and mental function. Ind J Med Res 1989;90:163-74.
  9. Soemantri AG, Pollitt E, Kim I. Iron deficiency anemia and educational achievement. Am J Clin Nutr 1985;42: 1221 -8.
  10. Pollitt E, Hathirat P. Kotchabhakdi NJ, Missell L, Valyasevi A. Iron deficiency and educational achievement in Thailand. Am J Clin Nutr 1989;50:687-97.
  11. Grantham-McGregor S. A review of studies of the effect of severe malnutrition on mental development. J Nutr 1995;125:2233S-8S.
  12. Pollitt E, Lewis NL, Garza C, Shulman R. Fasting and cognitive function. J Psychiat Res 1983;17:164-74.
  13. Simeon DT, Grantham-McGregor S. Effects of missing breakfast on the cognitive functions of school children of differing nutritional status. Am J Clin Nutr 1989;49: 64653.
  14. Carroll J. A model of school learning. Teachers Coll Rec 1963;64:723-44.
  15. Alexander KL, Entwisle DR, Dauber SL. First-grade classroom behavior: its short- and long-term consequences for school performance. Child Dev 1993;64 801 -14.
  16. Laird DA, Levitan M, Wilson VA. Nervousness in school children as related to hunger and diet. Med Rec 1931;134:494-9.
  17. Keister ME. Relation of mid-morning feeding to behavior of nursery school children. J Am Diet Assoc 1950; 26:25-9.
  18. Benton E, Brett V, Brain PF. Glucose improves attention and reaction to frustration. Bio Psychol 1987;24:95-100
  19. Galler JR, Ramsey F. Solimano G. Lowell WE. The influence of early malnutrition on subsequent behavioral development. II. Classroom behavior. J Am Acad Child Psych 1983;22:16-22.
  20. Richardson SA, Birch HG, Grabie E, Yoder K. The behavior of children in school who were severely malnourished in the first two years of life. J Health Soc Behav 1972;13:27684.
  21. Sigman M, Neumann C, Jansen AAJ, Bwibo N. Cognitive abilities of Kenyan children in relation to nutrition, family characteristics, and education. Child Dev 1989; 60:1 463-74.
  22. Wachs TD, Bishry Z. Moussa W. Yunis F. McCabe G. Harrison G. Sweifi E, Kirksey A, Galal O. Jerome N. Shaheen F. Nutritional intake and context as predictors of cognition and adaptive behavior of Egyptian school age children. Int J Behav Dev 1995;18:425-50.
  23. McDonald MA, Sigman M, Espinosa MP, Neumann CG. Impact of a temporary food shortage on children and their mothers. Child Dev 1994; 65:404-15.
  24. Chandler AK, Walker SP, Connolly K, Grantham-McGregor SM. School breakfast improves cognitive function in undernourished Jamaican children. J Nutr 1995;125:894900.
  25. Hamill PW, Drizd TA, Johnson CL, Reed RB, Roche AF. Growth curves for children, birth-18 years. Hyatsville, Md, USA: DHEW Pub. No. 78-1650, National Center for Health Statistics, 1977; Ser. 11:165,1-74.
  26. Must A, Dallal GE, Dietz WH. Reference data for obesity: 85th and 95th percentiles of body mass index (wt/ht2) and triceps skinfold thickness. Am J Clin Nutr 1 991 ;53:83946.
  27. Cohen J. A coefficient of agreement for nominal scales. Educ Psychol Meas 1960;20:3746.
  28. Himes JH. Reliability of anthropometric methods and replicate measurement. Am J Phys Anthropol 1989;79: 77-80.
  29. Clarke NMA, Grantham-McGregor SM, Powell C. Nutrition and health predictors of school failure in Jamaican children. Ecol Food Nutr 1991;26:1-11.
  30. Walker SP, Grantham-McGregor S. Himes JH, Williams S. Bennett F. Nutritional and health determinants of school failure and dropout in adolescent girls in Kingston, Jamaica. Research Report Series No. 1. Washington, DC: International Center for Research on Women, 1994.
  31. Velleman PF, Hoaglin DC. Applications, basics and computing of exploratory data analysis. Boston, Mass, USA: Duxbury Press, 1981.
  32. Cohen J. Statistical power analysis for the behavioral sciences. Hillsdale, NJ, USA: Lawrence Erlbaum Associates, 1988:273406.
  33. Harbison RW, Hanuskek EA. Educational performance of the poor. Lessons from rural northeast Brazil. New York: Oxford University Press, 1992.
  34. Rosenfield P. Lamber NM, Black A. Desk arrangement effects on pupils' classroom behavior. J Ed Psych 1985; 77:101-8.
  35. Weinstein CS. The physical environment of the school. A review of the research. Rev Ed Res 1979;49:577610.

Contents - Previous - Next