We propose that students experience "autonomy dissatisfaction" when the learning environment is indifferent to their psychological need for autonomy. We hypothesized that (a) students could distinguish this newly proposed need state from both autonomy satisfaction and autonomy frustration, (b) autonomy dissatisfaction would explain unique and rather substantial variance in students' classroom disengagement, and (c) a full understanding of the psychological need for autonomy necessitates expanding the current emphasis from two need states (satisfaction, frustration) to three (dissatisfaction). In the experimental condition, 20 secondary-school physical education (PE) teachers learned how to teach in an autonomy-supportive way; in the control condition, 17 PE teachers taught using "practice as usual." Their 2,669 students (1,180 females, 1,489 males) self-reported their autonomy satisfaction, autonomy dissatisfaction, autonomy frustration, engagement, and disengagement throughout a semester. Objective raters scored the manipulation check (teachers' autonomy-supportive instructional behaviors) and the engagement-disengagement outcome measure. Autonomy dissatisfaction longitudinally increased in the control group and longitudinally decreased in the experimental group. Most importantly, intervention-enabled decreases in autonomy dissatisfaction decreased students' end-of-semester disengagement, even after controlling for midsemester changes in autonomy satisfaction and autonomy frustration. We discuss the theoretical and practical benefits of adding autonomy dissatisfaction to the self-determination theory explanatory framework.