This study examined the relationships between cognitive development, language acquisition, and socioeconomic class, as manifested in children's written language. Four classes of sixth-grade students participated in the investigation, a control class and an experimental class from each of two socioeconomic environments (low and high). The treatment consisted of 12 science-inquiry films and six discussion sessions, during which children were taught to generate hypotheses of higher quality. At the end of each film session the students wrote as many hypotheses as they could in a rigorously controlled 12-minute period. Students in the control group responded to three additional film sessions. The collected papers were scored for hypothesis quality and syntactic complexity. Results indicated that socioeconomic level was not a variable in the relationship between syntactic complexity and hypothesis quality; scores for these two variables were positively correlated for control and treatment groups at each socioeconomic level. Both hypothesis quality and syntactic complexity were higher for the groups given instruction in the formation of hypotheses. Tables of findings are included. (AA)