Learning genre from examples: The effects of number of examples and scaffolding on student understanding and writing

This study investigated (1) whether students can learn unfamiliar genre when instruction includes conditions found to support learning from examples across domains, (2) how instructional conditions impact learning, and (3) whether patterns in knowledge gained when learning from examples in other domains also characterize what students learned about this genre. Over three days in four 10th grade English classes, students learned the sestina, a seven stanza poetic form in which each stanza ends with the same set of words rotating in a particular order. All students were given the same core instructional treatment which included an expert sestina, a description of critical genre features, and tasks prompting active feature analysis and composition. Half the students also received an additional two examples and half received additional directions scaffolding analysis and use of particular features. The four conditions were Single, Single Scaffolded, Multiple, Multiple Scaffolded. At pretest, students discussed what they noticed about an expert sestina. During instruction, students composed an original sestina. At posttest, students critiqued their own sestina, wrote directions for composing a sestina, and analyzed an unusual expert sestina. All data were analyzed for mention or use of 25 critical sestina features, ten regarding form and fifteen regarding craft.; Students knew little or nothing about the sestina before instruction. There were no differences by condition or by ability in pretest genre knowledge. All students across condition and ability knew a significantly greater number of sestina features after instruction as indicated by a comparison of pretest, posttest, and poem performance. Students could use the sestina features in their original poems and discuss these features explicitly on the posttest. Students used a greater number of sestina features in their poems than they mentioned on their posttests and learned surface form features more easily than deeper craft features. There were no significant differences between conditions. The large gains in genre knowledge by all suggest that the core instructional treatment supported learning. There were qualitative differences in sestina knowledge reflective of the example(s) students studied. Students wrote like the examples they studied. Use of multiple assessments captured variation and depth of genre knowledge.