Background: Specificity and transfer of learning have been examined in experimental studies. However, their findings may not be relevant to practitioners because of the difference between the experiment conditions and teaching situations. This case study investigates the theoretical issue of specificity vs. transfer of learning by conducting balance training with real-life (functional) tasks and evaluating the teaching outcomes through teaching observation of functional task performance and on the performance of balance tests. Purpose: To describe how a case formulation approach to problem-driven intervention for common balance problems in children can be conducted on the bases of eclectic theories, and to explore the reasons why observed learning outcomes of balance training can be, or can not be reflected on the performance of balance test items. Participants and setting: Two children who met the diagnostic criteria of developmental coordination disorder participated in this study. They were individually taught by student teachers at a teaching and research laboratory for 17 weeks over a 28-week period.Research design: We used a mix-method multiple case study design with two cases to test the alternative hypotheses, specificity vs. transfer of learning by theoretical generalization, in which each case is tested against the hypotheses. Data collection: The effects of balance training were assessed with the Movement Assessment Battery for Children (MABC) and through teaching observation of functional task performance. Data analysis: The analytical method of pattern matching was used to examine the match between the improvement patterns shown by balance test tasks and real-life balance tasks. Findings: Training resulted in an improvement of the balance skills that corresponded with the sets of tasks and environments employed for the specific balance training in one case. In the other case improved performance occurred not only on the trained balance tasks, but also on the performance test balance tasks in the MABC. Conclusions: Transfer and specificity of learning can be explained by the similarities between teaching tasks and test items, the directions of transfers, and the possible contributions of non-balance factors.