The authors examined parent-child interactions among three groups: selectively mute, anxious, and non-anxious children in different contexts. The relation between parental control (granting autonomy and high power remarks), child factors (i.e., age, anxiety, verbal participation), and parent anxiety was investigated. Parental control varied by context but parents of children with SM were more controlling than parents in the comparison groups in all contexts. Regression analyses indicated that child and parent anxiety predicted parental control, with increased anxiety associated with increased control. Older child age predicted less parent control. Group categorization moderated the relation between age and high power remarks, such that age was not a significant predictor for children with SM. Finally child-initiated speaking predicted high power remarks over and above other variables. These results support previous theories that parents take over for their children when they fail to meet performance demands, especially when the child or parent is anxious.