Related research suggests that children may show some simple understanding of event covariations by the early elementary school years. The present experiments use a rule analysis methodology to investigate covariation judgments of children in this age range. In Experiment 1, children in second, third and fourth grade judged covariations on 12 different covariation problems. Children's performance patterns on the problem set showed an increase in the use of systematic judgment strategies in this age range. Systematic rule users most commonly compared contingency table cells "a" and "b" in judging the event covariations. In Experiment 2, a training paradigm was employed to investigate possible origins of systematic rule use. First and second grade unsystematic, strategy O and cell-a children were either directed to attend to cells "a" and "b" (Attention only), were additionally offered explicit instructions to note which of the two cells had more events (Attention-plus-more) or were given no training (control). Posttest performance showed that the Attention-plus-more condition was the only treatment to reliably elicit a-versus-b rule use. It is concluded that simple covariation judgment rules can be used by children in the early elementary school years. (Author)