Evaluation methods for child feeding projects in developing countries, final report : Chapt. IV, an analysis of pilot study survey data in three countries

THIS CHAPTER OF THE FINAL REPORT ON THE STUDY CONTAINS A DETAILED DISCUSSION OF THE DESIGN OF THE ANALYSIS, THE VARIABLES ANALYZED, AND THE RESULTS OF THE ANALYSIS. ANALYSES OF DATA USING THE INDIVIDUAL CHILD AS A UNIT OF OBSERVATION WERE GENERALLY INCONCLUSIVE. THE INCONCLUSIVE FINDINGS ARE BELIEVED TO BE THE RESULT OF IMPRECISION IN MEASUREMENT OF SURROGATES FOR PROGRAM EXPOSURE, DEFINED AS SIZE OF RATION CONSUMED AND LENGTH OF PARTICIPATION IN A SUPPLEMENTARY FEEDING PROGRAM. SUCH MEASUREMENT PROBLEMS ARE CHARACTERISTIC OF THE CROSS-SECTIONAL APPROACH. MATERNAL-CHILD HEALTH (MCH) PROGRAMS APPEAR, IN GENERAL, TO BE MORE EFFECTIVE IN MEETING COMMUNITY NUTRITIONAL NEEDS THAN SCHOOL FEEDING PROGRAMS. AS EXPECTED, MCH PROGRAMS REACH POORER FAMILIES, MOTHERS, AND NEWLY WEANED CHILDREN AT TIMES WHEN THEIR NUTRITIONAL NEEDS ARE GREATEST. HOWEVER, COSTS PER CHILD FOR MCH PROGRAMS ARE HIGHER THAN FOR SCHOOL FEEDING PROGRAMS. GIVEN THE PREVAILING SYSTEM FOR GIVING OUT MCH RATIONS, MUCH OF THE SUPPLEMENTAL FOOD IS CONSUMED BY PERSONS OTHER THAN THE TARGETED PREGNANT AND LACTATING MOTHERS AND PRE-WEANED OR POST-WEANED CHILDREN. SCHOOL FEEDING PROGRAMS APPEAR TO BE MOST EFFECTIVE WHEN THEY PROVIDE A HIGH PROTEIN SUPPLEMENT TO SUSTAIN GROWTH. A WEIGHT DISPERSION MEASURE DEVELOPED IN THIS STUDY OFFERS ADVANTAGES OVER "PERCENT OF STANDARD WEIGHT-FOR-HEIGHT" AS A MEANS OF MEASURING THE NUTRITIONAL IMPACT OF A FEEDING PROGRAM.