Objectives: The present study tested the potential for manipulating adolescents' goals and motives for participation in a school physical education (PE) lesson, and explored the subsequent outcomes on participant experience. Design: A cluster randomized controlled design was used to compare outcomes of four experimentally manipulated PE class conditions alongside a control group. Method: Twenty-four classes comprising 592 students (M age = 13.74 years) were randomized to undertake one of four experimental conditions, or a. control condition, during a fitness-based circuits class during a usual PE lesson. The experimental conditions comprised an autonomy-supportive or controlling climate, each with an intrinsic (health and fitness) or extrinsic (looking good to others) goal focus. The control condition comprised a neutral climate with no goal focus. The effect of experimental condition on motivational, affective, and intentional outcomes was analysed using hierarchical linear modelling. Results: Class-level effects explained less than 10% of variance of study outcomes, suggesting that the impact of lesson manipulations was limited. Where intervention effects were significant, these were contrary to hypotheses guided by self-determination theory (SDT); participants perceived greater lesson value and formed stronger future intentions in the controlling, extrinsic goal focused condition. However, at the individual-level, findings were in line with SDT, in that perceptions of autonomy support and an intrinsic goal focus predicted positive lesson-related outcomes (i.e., motivation, effort, enjoyment, value, exercise-induced affect) and future intention to exercise (Total R-2 values = .39 to .75). Conclusions: The findings highlight the practical challenges of manipulating lesson climates in ecological PE settings. (C) 2012 Elsevier Ltd. All rights reserved.