A case-control study was conducted to know the magnitude of the effect of main risk factors for short stature of pre-school children in a city in São Paulo State, in 1995. An anthropometric survey carried out with all children (1201) attending public pre-school classes was used to select the case and the control groups. It was selected a random sample of 165 children among those with stature/age <= -1 Z score (case) and another sample of 165 among those with stature/age > or = +1 Z score (control). The mothers or persons responsible for these children were interviewed in order to obtain information on demographic, maternal and socio-economic variables. The multiple logistic hierarchical analysis showed the following variables as associated with pre school children's short stature: mother's educational level (OR = 2,1; CI: 1,1-3,8); per capita family income <= 0,5 SM (OR = 3,4; CI: 1,5-8,0); number of persons in the house > or = 6 (OR = 3,7; CI: 1,5-9,0); number of domestic equipment <= 1 (OR = 4,4; CI: 1,8-10,7); birth length < 48 (OR = 7,4; CI: 2,3-23,7), mother's stature <= 156,6 (OR = 5,9; CI: 3,1-11,0) and father's stature (OR = 4,2; CI: 2,1-8,6). We was found that even in a population of preschool children without nutritional deficiency (as measured by the usual anthropometric index) it is possible to observe the effect of socio economic variables in children stature.