Writing to Learn in Science: Effects on Fourth-Grade Students' Understanding of Balance

Abstract

In this study, 69 grade 4 students were randomly assigned to a writing to learn (n = 23), comparison (n = 23), or control (n = 23) condition. In treatment and comparison, students made predictions and recorded results for 30 balance trials which involved different configurations of weights on an equal arm balance beam. Treatment students wrote 4 short answer responses about what they were learning during the trials and wrote an extended response after the trials about what they had learned about balance. To control for writing time, comparison students wrote 4 short answer responses about their favorite parts of the trials and wrote an extended response about their favorite parts of science class. Students in the control condition (n = 23) participated in business as usual classroom instruction. There were no statistically significant differences between treatment and comparison conditions on a 20-item balance knowledge posttest. However, students in the control condition outperformed students in the treatment (ES = 0.89) and comparison (ES = 1.05) on posttest questions at the lowest level of balance understanding (i.e., level 1). For balance knowledge questions at the highest levels, levels 3 and 4, the treatment condition (ESs = 1.42 and 0.94, respectively) and the comparison condition (ESs = 1.62 and 1.37, respectively) outperformed the control. No statistically significant differences were found between conditions for total words written and level of balance understanding on a posttest extended writing prompt. Implications of these findings and directions for future research on writing to learn are discussed.