Discriminant function analysis was used to understand the combination of socioeconomic variables in various forms of malnutrition in preschool children. Data available (1976) from two sets viz., 4000 children of rural and urban areas of Hyderabad (in south India) and 215 children of urban and rural areas of Jaipur (in north India) were used for the study. The prevalence rates of various forms of malnutrition in children were different by variations in socioeconomic status and sex (P less than 0.05). Discriminant function analysis of all socioeconomic indicators showed that educational levels and occupational status of parents and percaput income were most important for differentiation of children who were normals and undernourished. Occupational status of the father and the child's calorie intake or duration of lactation improved the differentiation of normals from those who were under-nourished. Higher prevalence of malnutrition in female children than in the male children was found due to variations in per capita income and levels of maternal education. Closely associated best set of factors could be traced using the discriminant function analysis.