This study investigated whether the amount and timing of group-based childcare between birth and 51 months were predictive of cognitive development at 51 months, taking into account other non-parental childcare, demographic characteristics, cognitive development at 18 months, sensitive parenting and a stimulating home environment. Children's (N=978) cognitive development was assessed at 51 months with four subscales of the British Ability Scales: two verbal and two non-verbal. Mothers were interviewed and observed at 3, 10, 18, and 36 months and the quality of group care was assessed at 10, 18, and 36 months (N=239) if it was used for [greater than or equal to] 12 hours per week. Age of starting in group care and amount were highly associated (r = -.75). Multiple regressions indicated that, controlling for other factors, higher cognitive development and particularly non-verbal ability was associated with more hours per week in group care from 0 to 51 months, or an earlier start, or group care before age 2. Nevertheless, the majority of variance was explained by other predictors: sex (girl), higher cognitive development at 18 months, older mother, first language English, mother of white ethnic background, with more qualifications, higher family social class, more maternal responsivity at 10 months and a more stimulating home learning environment (HLE) at 36 months. Hours per week in relative care or home-based care were not significant predictors of cognitive scores. For the smaller relatively advantaged sample who had group care quality information (N=239), quality was a marginal predictor of better cognitive development but age of starting group care was not. Most variance was explained by 18 month cognitive development, maternal education, and family social class. (author abstract)