The study investigated the relative effectiveness of two rule-governed approaches to teaching children how to control variables. In the elaborative-rules treatment students were given a mnemonic to stimulate recall of three elements of a well-designed experiment: causes, effects, and controls. In the organized-rules treatment students were given a well-organized set of six rules that provided detailed direction with respect to the same three elements of an experiment. Intact classes of grade-5 students (n = 283 students) were randomly assigned to two treatments and a control condition consisting of existing teacher practice; after four weeks control-group classes were reassigned to the rules treatments. Student achievement was observed prior to the implementation of conditions, after instruction, and three months later. Both rules treatment outperformed traditional instruction. The organized-rules method was slightly more effective overall, but the elaborative-rules approach had a slightly greater benefit for two groups which tend to be disadvantaged in science achievement: females and field-dependent students.