The PIs in this project from Temple University are examining the cognitive mechanisms that undergird two scaffolding strategies to improve student comprehension of diagrams. They are exploring the types of inferences that learners draw when they use self-explanation or sketching to explain visual representations in algebra and in biology. Working with middle school students, they are testing five comparisons of student performances on problems in those domains. They are examining the mechanisms underlying the success of these two interventions alone and in combination for mathematics and science problems to determine whether the two mechanisms are redundant or complementary. They are examining the nature of the inferences that are prompted by the two techniques and examining the differences in the ways that students with higher spatial ability use the strategies. They follow the performances of the students in a second year on other tasks using similar strategies to examine the how student sketching and self-explanation skills relate to long-term science and math learning.
Three groups of participants are completing individual computer-based sessions that include measures of spatial abilities, working memory, motivation for drawing and spatial tasks, and solving a series of math and science problems using sketching and/or self-explanation. Student performance on the problems are analyzed in terms of the pretest measures. One group of 200 students is being studied longitudinally in 6th through 8th grades to understand changes over time in the reasons why sketching and self-explanation are effective. A second group of 200 8th graders complete the measures in the first year of the project to provide a cross-sectional age comparison. A new group of 6th graders are being audiotaped while solving the problems in the 2nd year of the project in order to understand cognitive processes in finer-level detail. Analyses include cross-sectional t tests, regression models, longitudinal growth curve modeling, and analyses of verbal protocols from the audiotaped sessions. Research activities include developing low- and high spatial math and science problems and a coding scheme for the many different types of inferences. Three doctoral students are trained over the course of the project in school-based research, collection of verbal protocol data, coding, and data analysis. The team consists of experienced psychology researchers, with an expert advisory board providing an annual evaluation.
The products of the research are the low- and high-spatial problem sets, inference coding scheme, and scholarly papers reporting the results of the research. In addition to 600 participating students, the math and science teachers of those students are involved in summer work assisting with coding the student work. The project has an impact on teachers' use of self-explanation and/or sketching at many levels--from elementary through graduate education. The findings help point teachers towards when and with whom to use sketching and/or self-explanation, which leads to better learning in science and math.
