Central to argument are evidence-based claims, requiring coordination of a claim with evidence bearing on it. We advocate a dialogic approach to developing argument skills and in the work reported here examine the further scaffold of prompts that exemplify functions of evidence in relation to a claim. This scaffold was successful in accelerating the prevalence of evidence-based claims in essays of low-performing middle schoolers compared to participants in the same year-long dialog-based intervention who received no or a limited form of evidence prompts and compared to previous samples engaged in a nondialogic curriculum. An experimental group achieved a proportion of evidence-based claims above 50% by the end of the year, transferring their newly developing skill from one topic to another. The use of different types of evidence emerged in a sequence corresponding to the cognitive demands they posed. Students first used support-own evidence. They used weaken-other evidence increasingly over time, but the two evidence types inconsistent with their position (support-other and weaken-own) showed lesser and later gains. Supporting a dialogic approach, qualitative data showed that evidence use occurred most readily in dialogs; then in individual writing on the same topic; and to a more limited extent in essays on a new, unstudied topic.
Stored only when both papers are in the corpus — 74 citations exist upstream.