Differences on 13 child language variables were investigated for 12 sels of 24-ycar-old twins and 24 singletons, matched lor age and socioeconomic stains. Two-hour spontaneous speech records were obtained in the home. A regression analysis assessed the contribution of the biological (perinatal) and maternal speech variables to the child language measures. Rate of mother's speech to child was the most important predictor of child language performance. Singletons talked more, addressed more utterances to iheir mothers, and combined phrases into sentence (subject + verb) form to a greater extent than did the twins. Environmental factors appear to be influential in encouraging use of verbal expression beyond acquisition of phrase competence. Biological factors make a smaller contribution.Most measures of twin language and intellectual development show twins progressing at a slower rate than singletons (Day, 1932;Koch, 1966;Mittler, 1971;Breland. 1974). Some of the difference may be due to the biological effects of twin pregnancies and births, which may entail competition for space and nutrition in utero and always result in lowered birthweight (McKeown & Record, 1952;Naeye, 1964; Babson & Bramhall, 1969;Howard & Brown, 1970; Wilson, Note 1). However, some of the difference in language may also be due to the parental environment, and particularly parental verbal stimulation, which twins, as opposed to singletons, have to share (for the importance of linguistic environment cf. Cross, 1977;Snow, 1972 Snow, , 1977 Seitz& Stewart, 1975; Whitehurst, Note 2). Record, McKeown and Edwards (1970) found that twins whose partners had died before one month of age had average verbal reasoning quotients at age 11, whereas the verbal reasoning scores of full twin sets were below average. They concluded from this and from their negative findings regarding birth factors that lower verbal intelligence in twins was not due to prenatal or perinatal handicapping conditions, but to post-natal and environmental experiences.None of these authors, however, examined directly the way in which twins differ from singletons in their day-to-day experiences in the home. Lytton, Conway, and Sauve (1977) did just this, and obtained first-hand records of twins', singletons', and their respective parents' speech in a global way. The study reported here examined the structure and complexity of twins', singletons', and their respective mothers' speech in detail, Requests for reprints should be sent to

No abstract

No abstract

This study explored the likelihood that current inconsistencies in findings relating to the programming variables of step size (SS), response mode (RM) and knowledge of results (KR) might stem from the tendency of past research to investigate these variables in isolation from each other. Such an approach fails to make provision for the observation of possible interactions. Accordingly, four levels of KR (O, 50% variable-ratio, 100%, 100% plus incentive), three levels of RM (overt, covert, covert plus random overt) and two levels of SS (large, small) were incorporated into a factorial design. Three intact classrooms were randomly assigned to each of the twenty-four treatment conditions. A linear program with content structured to teach certain concepts pertaining to earth-sun relationships served as the learning task. The treatment combinations were evaluated for their effects upon immediate learning and delayed retention. Immediate learning was maximized by treatments that coupled a small step program with an overt response mode. KR was generally not an important contributor to immediate learning. Analysis of immediate learning data also revealed significant SS×RM and SS×KR interactions. There were no differences, however, among treatments in terms of facilitating delayed retention of content. It was concluded that a reappraisal is required of the pivotal assumption that programmed instruction's effectiveness derives from the explicit provision of KR.