|Early Supplementary Feeding and Cognition (Society for Research in Child Development, 1993, 123 pages)|
Ernesto Pollitt and Kathleen S. Gorman
The Commentaries by Theodore D. Wachs and Nevin S. Scrimshaw complement the work presented here and, for the most part, represent a natural extension of the Monograph discussion. Scrimshaw's discussion helps frame the current findings within a historical perspective regarding, the study of malnutrition and behavior. His knowledge of the history of this topic is perhaps one of the most extensive of anyone in the field. We do take issue, however, with the rather pessimistic view evident in some of his remarks. Our understanding of human behavioral development, plasticity, and malleability leads us to a more optimistic view whereby nutritional insults are sensitive to interventions. In our view, with the exception of iodine deficiency, none of the nutritional deficiencies that represent major public health problems for a population result in an irreversible impairment of cognitive function.
The comments by Wachs flow naturally from the work in which he has been involved over the past years, in relation to both developmental theory and the significant role of malnutrition on development. At the same time, Wachs raises several issues that we feel merit further discussion. First, it is true that our measures of information processing were less stable than those of abilities and achievement. We agree that no definitive conclusions can be drawn on whether the absence of positive results regarding the nutritional intervention on information-processing variables is a result of the lack of intervention effects or the measures used.
The issue of sex differences is an interesting one. Originally, all analyses were run separately by sex. Given a very similar pattern of results, we combined the sample and tested for sex x treatment interactions. Finding no significant interactions, we then removed the interactive term and included sex as a main effect only. In a separate publication (Gorman & Pollitt, 1992), we have reported that males attain a significantly higher grade than females but that the age at which children start school is equivalent for both sexes. As the focus of the analyses in the Monograph was on the effects of the treatment, we did not explore all possible potential interactions. Following Ted Wach's comments, we have gone back and looked at the potential three-way interactions (treatment x sex x maximum grade attained) and found only limited evidence for differential effects based on gender. Only in two cases were the interactions significant (i.e., literacy and knowledge), and, when analyses were separated by gender, only for the knowledge test did the interactions remain significant. Similar to the results of the two-way interactions for all subjects combined, among females grade attainment was positively related to performance on the knowledge test in Atole villages, whereas there was no such association in the Fresco villages. Among males, this relation was not statistically significant.
The comments made by Wachs regarding double buffering raise some interesting ideas that require further clarification. We understand the results to show that, first, the Atole treatment was a buffer against the adverse effects of low SES and, second, high SES protected the subjects against the adverse effects of poor diet. Double buffering suggests to us an additive effect of Atole and SES whereby Atole subjects from the high-SES families would have outperformed the high-SES Fresco subjects. This was not the case. The results of the schooling interactions do suggest a double buffering whereby schooling and treatment combined were associated with the highest performance. Again, we would maintain that the different patterns of interactions reflect differences in variability of opportunity provided by SES and schooling.
Finally, we agree with the clarifications and additions of the model proposed by Wachs, which include the important cultural and ecological context in which behavior occurs. We would also point out that growth and activity are not necessarily independent effects and that caregiver behavior and exploration are not independent processes. The process is complex, and one could expect multiple points of interaction.
In terms of policy implications, Wachs argues that the data are strongest in supporting interventions for low-socioeconomic-status children, and we fully agree. However, not all children who benefited from the treatment were necessarily from low-SES backgrounds. Therefore, by targeting only low-SES children, the potential benefits of the nutritional intervention for other children who remain in school would be lost. These broad-range effects in the context of a rural and economically impoverished community need to be accounted for in arguments that favor feeding only the poorest of the poor. This would be an enormous disservice to children everywhere. Interventions must be multifocal, ongoing, and available to all children in need. In this sense, we agree with Wachs in asserting the importance of nutritional aid, economic aid, and keeping all (our emphasis) children in school for as long as possible.
Gorman, K., & Pollitt, E. (1992). School efficiency in rural Guatemala. International Review of Education, 38(5), 519 - 534.