for advice about meta-analysis and statistical analyses; Adele Gottfried and Vivian Gadsden for their input regarding interpretation of findings; and Stuart Yager for consultation regarding his contributions to the literature. Cynthia A. Rohrbeck and Marika D. Ginsburg-Block contributed equally to this study.

No abstract

Meta-analysis was used to examine social, self-concept, and behavioral effects of peer-assisted learning (PAL) interventions with elementary school students. An electronic search of PsycINFO and ERIC databases resulted in 36 relevant PAL studies. Overall, effect sizes were small to moderate across the 3 outcome variable domains. Both social and self-concept outcomes were positively correlated with academic outcomes. Specific PAL components-student autonomy, individualized evaluation, structured student roles, interdependent group rewards, and same-gender grouping-were related to effect sizes. PAL interventions were more effective for low-income versus higher income, urban versus suburbanrural, minority versus nonminority, and Grades 1-3 students versus Grades 4 -6 students. Results suggest that PAL interventions that focus on academics can also improve social and self-concept outcomes.

The effects of 2 instructional methods, problem solving and peer collaboration, were evaluated for enhancing mathematics achievement, academic motivation, and self-concept of 104 low-achieving 3rd and 4th graders. Students were assigned randomly to 1 of 4 conditions: control, problem solving, peer collaboration, and problem solving + peer collaboration. Students in all conditions met twice weekly for 30-min mathematics sessions over a 7-week period. Results indicate that problem-solving students performed significantly higher man their counterparts who did not receive problem solving on measures of computation and word problems and reported higher levels of academic motivation, academic self-concept, and social competence. Students who participated in peer collaboration scored higher on measures of computation and word problems and reported higher levels of academic motivation and social competence than did students who did not participate in peer collaboration. No significant interaction effect was found.For the United States to succeed in the 21st century, youth must be prepared to achieve in the areas of mathematics and science (U.S. Department of Education, 1991). However, national statistics indicate that students, particularly those living in low-income urban areas, are not performing adequately in these subjects. In 1992, the National Assessment of Educational Progress revealed that one third of the fourth and eighth graders performed below basic levels of proficiency in mathematics. Evaluations of students living in low-income urban settings indicated that more than two thirds failed to evidence basic levels of mathematics achievement (Children's Defense Fund, 1994).Efforts by educators to improve educational outcomes for diverse learners have traditionally focused on either design-

The authors wish to thank Gwen Davis and Barbara Schaefer for their important assistance in this research effort and Tamara Marder and Jennifer Eaton for their able assistance in coding videotapes. Also special thanks to our elementary school partners who helped us design and carry out this study. These educators include prinicipal Mr. Manns and participating teachers Barbara Felzer, Barbara

This longitudinal study examined the predictive validity of the MacArthur Communicative Developmental Inventories-Short Form (CDI-SF), a parent report questionnaire about children's language development (Fenson, Pethick, Renda, Cox, Dale & Reznick, 2000). Data were first gathered from parents on the CDI-SF vocabulary scores for seventy-six children (mean age=1 ; 10). Four years later (mean age=6 ; 1), children were assessed on language outcomes (expressive vocabulary, syntax, semantics and pragmatics) and code-related skills, including phonemic awareness, word recognition and decoding skills. Hierarchical regression analyses revealed that early expressive vocabulary accounted for 17% of the variance in picture vocabulary, 11% of the variance in syntax, and 7% of the variance in semantics, while not accounting for any variance in pragmatics in kindergarten. CDI-SF scores did not predict code-related skills in kindergarten. The importance of early vocabulary skills for later language development and CDI-SF as a valuable research tool are discussed.

No abstract

No abstract