Experimental research on strategy instruction for beginning writers has only recently begun. This study investigated the role of self-regulation instruction in Grade 1 strategy learning. In a pretest-post-test quasi-experiment, 120 Grade 1 students participated in a unit of study on personal narrative in one of three conditions: (1) Self-Regulated Strategy Development (SRSD), which included instruction in goal-setting, strategy steps, coping, self-monitoring, and self-reinforcement; (2) strategy instruction only (SO), which focused on the goal and steps of the strategy; (3) a control condition. Students in both strategy conditions, relative to the control, made large, statistically significant gains in text quality, word count, story features, and self-regulation knowledge. The SRSD condition resulted in greater self-regulation knowledge than the SO condition, which resulted in greater knowledge than the control. Pretest text quality did not interact significantly with condition. The effect of instruction on text quality was largely mediated by post-test self-regulation knowledge.