We illustrate this approach by examining the relationships of training loads with a performance-related state (鈥榮elf-efficacy鈥? and the role of potential mediating PBS variables (鈥榝atigue/lack of energy鈥?and 鈥榖eing in shape鈥? in explaining these relationships in an elite triathlete across time.
Self-reports of PBS states (twice weekly) and training data were collected over 137 days. Using DLMs and mediating variable analysis, direct (unmediated) and indirect (mediated) short-term associations of training load with 鈥榮elf-efficacy鈥?were examined.
In this triathlete, we found evidence for positive effects of training on 鈥榮elf-efficacy鈥? which were partly explained by feelings of 鈥榖eing in shape鈥?and suppressed by feelings of 鈥榝atigue/lack of energy鈥? Changes in the relationship between lagged training load and 鈥榝atigue/lack of energy鈥?were observed across time and were particularly pronounced in temporal proximity of an injury.
Strengths of the presented approach are its dynamic nature enabling the observation of changes occurring over time, use of statistical inference rather than visual data interpretation, and quantification of mediating effects to identify potential pathways of intervention. Additionally, the DLM method can identify complex nonlinear associations by examining correspondence between changes in levels of predictors and changes in magnitude and direction of predictor-outcome associations.