Recent research suggests that using naturalistic teaching paradigms leads to therapeutic gains in clinic settings for children with autism and related disorders. More recent studies are demonstrating that implementing these strategies within a parent training format may produce collateral effects in other areas of family life. The present experiment assessed collateral effects of two very different parent training paradigms during unstructured dinnertime interactions in the family setting. One paradigm focused on teaching individual target behaviors (ITB) serially, and the other focused on a recently developed naturalistic paradigm that teaches the pivotal responses (PRT) of motivation and responsivity to multiple cues. Two groups of families were randomly assigned to each of the parent training conditions. Pretraining and post-parent-training videotapes of dinnertime interactions were scored in a random order across four interactional scales (level of happiness, interest, stress, and style of communication). Results obtained for the four interactional scales showed that the families in both conditions initially scored in the neutral range, and the ITB training paradigm produced no significant influence on the interactions from pretraining to posttraining. In contrast, however the PRT parent training paradigm resulted in the families showing positive interactions on all four scales, with the parent-child interactions rated as happier, the parents more interested in the interaction, the interaction less stressful, and the communication style as more positive.