LAUSR

Abstract In this study we sought to identify profiles of talk during Head Start preschool mealtime conversations involving teachers and students. Videos of 44 Head Start classrooms’ lunch interactions were analyzed for the ratio of teacher–child talk and amount of academic vocabulary, and then coded for instances of academic/food, social/personal, and management talk to highlight the degree of hybridity of talk within this unique setting. Cluster analysis revealed four distinct patterns of teacher–child mealtime interactions in 44 Head Start preschool classrooms: classroom discourse, home discourse, hybrid-low, and hybrid-high. Multilevel models further demonstrated a relationship among these clusters of teacher–child interactions and children's end-of-year expressive vocabulary scores controlling for ratio of teacher–child talk and pre-test scores. Children in classrooms displaying a hybrid style of mealtime discourse made the greatest gains on measures of expressive vocabulary in contrast to their peers in classrooms displaying other discourse styles.