The aims of this study were to determine whether scapular and humeral head position can predict the development of shoulder pain in swimmers, whether those predictors were applicable to non-swimmers and the annual rate of shoulder pain in adolescent swimmers and non-swimmers. Forty-six adolescent swimmers and 43 adolescent non-swimmers were examined prospectively with a questionnaire and anthropometric measures. The questionnaire examined demographic and training variables. Anthropometric measures examined the distances between the T7 spinous process and the inferior scapula (Inferior Kibler) and T3 spinous process and the medial spine of the scapula (Superior Kibler), humeral head position in relation to the acromion using palpation, BMI and chest width. Shoulder pain was re-assessed 12 months later by questionnaire. Shoulder pain in swimmers was best predicted by a larger BMI (OR = 1.48, P = 0.049), a smaller Inferior Kibler distance in abduction (e.g. OR = 0.90, P = 0.009) and a smaller horizontal distance between the anterior humeral head and the anterior acromion (OR = 0.76, P = 0.035). These variables were not significantly predictive of shoulder pain in non-swimmers. Annual prevalence of shoulder pain was 23.9% in swimmers and 30.8% in non-swimmers (χ(2) = 0.50, P = 0.478).