A difference in length or height growth is observed between populations in developing and developed countries and, within a country, between well-off and poor subpopulations. The aim of this community-based study was to identify possible risk factors for growth stunting in early life using the infancy-childhood-puberty (ICP) growth model. A periurban slum, a village and an urban slum, plus a local control group in Lahore, Pakistan, were investigated. In total, there were 425 infants born from September 1984 to March 1987 in the study. Children were followed from birth to 24 months and anthropometric data were recorded at each examination. A well-designed questionnaire was used to collect background information to identify risk factors for growth faltering over this period. The age at onset of the childhood component was determined individually using the ICP growth model. A late onset was defined as onset after 12 months. Univariate odds ratio analysis identified 20 significant risk factors for a delayed childhood onset in the total sample. Backward logistical analysis showed that only six variables were significant, i.e. small number of rooms in a household, episode(s) of diarrhoeal diseases, poor house structure, low body mass index, early end of breastfeeding and birth in warm season. Stratified analysis showed that episode(s) of diarrhoeal diseases and birth in warm season were the only significant risk factors in the two poor areas. This study also found that the impact of diarrhoeal diseases on length growth persisted after allowing for birth season. Seasonal variation in length growth was seen in the age at onset of the childhood component. General socioeconomic status was found to be a risk factor for growth faltering in early life. Diarrhoeal diseases and being born in a warm season are particular risk factors for growth faltering in early life.