Athlete self-talk is often measured by self-report, but seldom are self-reports compared to observed, audible self-talk. In this study, 28 tennis players attending a youth camp (Mage = 12 years) were observed in practice and match play for positive, negative, motivational, and instructional self-talk. They reported how often they used self-talk phrases (Automatic Self-Talk Questionnaire) and how much of their speech was aloud and inside their head (Self-Talk Use Questionnaire). We observed the same players on the court for overt self-talk use during practice and match play (Self-Talk and Gestures Rating Scale). In observation, players used more negative than positive self-talk, and they infrequently used instructional self-talk in matches and practice. Tennis athletes who reported more overt self-talk were observed to use more self-talk during tennis. In most cases, athlete self-report of positive and negative self-talk did not relate to observation of positive and negative self-talk. We argue that self-report for youth athletes needs to be confirmed by observation.