Objective: To evaluate whether children's agricultural work practices were associated with agricultural injury and to identify injury and work practice predictors. Design: Analyses were based on nested case-control data collected by the Regional Rural Injury Study-II (RRIS-II) surveillance study in 1999 and 2001 by computer-assisted telephone interviews. Subjects: Cases (n = 425) and controls (n = 1886) were persons younger than 20 years of age from Midwestern agricultural households. Those reporting agricultural injuries became cases; controls (no injury) were selected using incidence density sampling. Main outcome measures: Multivariate logistic regression was used to estimate the risks of injury associated with agricultural work, performing chores earlier than developmentally appropriate, hours worked per week, and number of chores performed. Results: Increased risks of injury were observed for children who performed chores 2-3 years younger than recommended, compared to being "age-appropriate" (odds ratio (OR) = 2.6, 95% confidence interval (CI) = 1.4-4.5); performed any agricultural work (3.9 (2.6-5.6)); performed seven to ten chores per month compared to one chore (2.2 (1.3-3.5)); and worked 11-30 or 31-40 h per week compared to 1-10 h (1.6 (1.2-2.1) and 2.2 (1.3-3.7), respectively). Decreased risks of injury were observed for non-working children compared to children performing what are commonly considered safe levels of agricultural work. Conclusions: This study demonstrated elevated risks of agricultural injury among children who perform developmentally inappropriate chores. Results suggest that the efficacy of age restrictions for preventing the occurrence of childhood agricultural injuries warrants further evaluation.