OBJECTIVE: To assess the relationships between shy-inhibited temperament in childhood and anxiety problems in early adolescence using a prospective, longitudinal data set from a large community sample. METHOD: Relationships between shyness ratings on age-appropriate temperament scales and anxiety problems were analyzed, looking both forward and backward in time from infancy to adolescence. RESULTS: Forty-two percent of children rated as shy on 6 or more occasions over 8 surveys in childhood had anxiety problems in adolescence, compared with 11% who were never shy. Persistence of shyness and its presence in middle childhood increased risk for anxiety. A highly reactive temperament added to shyness did not increase the risk for anxiety. Few children with an anxiety diagnosis in early adolescence had a history of shyness. CONCLUSIONS: Prediction from childhood shyness to adolescent anxiety disorder is modest but clinically meaningful in a community sample. However, most shy children did not develop an anxiety disorder and most adolescents with anxiety disorders had not been especially shy.