Careful what you say to yourself: Exploring self-talk and youth tennis performance via hierarchical linear modeling

Researchers have examined self-talk and performance in tennis matches, but the current understanding of speech-to-performance associations in tennis is hindered by conflicting evidence from self-report and observation, and by the use of relatively simple statistical analyses that fail to account for speech and performance events being nested within both points and individuals over time throughout a tennis match. In the current study, 28 youth tennis players (M_age = 12 years) were observed, and their speech and performance coded with the Self-Talk and Gestures Rating Scale. The relation between self-talk and performance was measured on the same point (K), as well as on the subsequent point (K+1). Bivariate (non-nested) analyses supported a concurrent, but not a predictive relation between self-talk and performance, consistent with prior work. With multivariate, hierarchical linear modeling that accounts for total points played in a set within person, and player behavior on the prior point, we found that positive and negative self-talk were strongly related to concurrent point performance, and to a less extent, to subsequent K+1 point performance. The occurrence of positive self-talk on point K appeared to increase the odds of winning the subsequent point compared to when positive self-talk talk did not appear, but the margin was small. Researchers should continue to observe self-talk in athletes, as well as examine multilevel speech-to-performance associations, especially when data are gathered within naturalistic settings in real time.