A mixed-type Bayesian-network approach combines ordinal questionnaire responses, categorical background information and continuous indicators without forcing them into a single data type. The method pairs a latent Gaussian copula with constraint-based and score-based structure learning, with a bootstrap variant to improve stability. In simulations it recovers network structure accurately. Applied to data from 164 female volleyball players in Italy, it reveals relationships among mental skills, personality and self-esteem, and supports scenario analyses of how changes in skills relate to other attributes.