ABSTRACT OBJECTIVE To investigate the occurrence of daytime sleepiness and associated sleep factors in a sample of elementary school students who attended school in the afternoon schedule. METHODS Sleep data from 363 Brazilian public school students (12.78 ± 1.36 years, 206 girls) were obtained by applying questionnaires in classrooms. All subjects attended school in the afternoon schedule, with classes starting between 1:00 and 1:20 p.m. Daytime sleepiness was assessed by the pediatric daytime sleepiness scale; sleep quality, by the mini-sleep questionnaire; and sleep patterns and chronotypes, by the Munich chronotype questionnaire. Scores equal to or greater than 15 pediatric daytime sleepiness scale points were considered as excessive daytime sleepiness. The predictive power of sleep variables on daytime sleepiness was evaluated by a multiple linear regression. RESULTS The subjects in the sample had an average time in bed greater than nine hours both on school days and on weekends. Nevertheless, 52.1% had an average pediatric daytime sleepiness scale score equal to or greater than 15 points, indicative of excessive daytime sleepiness. As for their quality of sleep, 41.1% had a very altered sleep. We observed, by a multiple linear regression, that quality of sleep (β = 0.417), chronotype (β = 0.174), mid-sleep on school days (β = 0.138), and time in bed (β = - 0.091) were all significant in predicting daytime sleepiness. CONCLUSION This study showed the occurrence of excessive daytime sleepiness in non-sleep deprived students who attended school in the afternoon. The worst quality of sleep and eveningness had a greater predictive power on daytime sleepiness than time in bed. Therefore, we must consider other factors in addition to sleep duration when planning interventions for daytime sleepiness.