Synergy Between Teacher Practices and Curricular Scaffolds to Support Students in Using Domain-Specific and Domain-General Knowledge in Writing Arguments to Explain Phenomena

Abstract

We investigated how 2 different curricular scaffolds (context-specific vs. generic), teacher instructional practices, and the interaction between these 2 types of support influenced students' learning of science content and their ability to write scientific arguments to explain phenomena. The context-specific scaffolds provided students with hints about the task and what content knowledge to use in or incorporate into their writing. The generic scaffolds supported students in understanding a general framework (i.e., claim, evidence, and reasoning) regardless of the content area or task. This study focused on an 8-week middle school chemistry curriculum that was enacted by 6 teachers with 578 students during the 2004–2005 school year. Analyses of identical pre- and posttests as well as videotapes of teacher enactments revealed that the curricular scaffolds and teacher instructional practices were synergistic in that the effect of the written curricular scaffolds depended on the teacher's enactment of the curriculum. The context-specific curricular scaffolds were more successful in supporting students in writing scientific arguments to explain phenomena, but only when teachers' enactments provided explicit domain-general support for the claim, evidence, and reasoning framework, suggesting the importance of both types of support in successful learning environments.

Notes

¹Because of our concern about Type I error, we first analyzed the teachers' data together to determine whether there was a main effect of scaffold type across all of the teachers. Because there were two different discrete dependent variables of interest (scaffold type and teacher), we could not use a traditional post hoc test in conjunction with the ANCOVA, such as a Tukey test or Bonferroni test, which are only designed for one dependent variable. Furthermore, we could not include either variable (scaffold type or teacher) as a covariate because they were not continuous. Ideally, we would have used multilevel modeling, such as hierarchical linear modeling, to analyze this effect, but our teacher sample size was too small.

²For all four ANCOVAs presented in this section, the interaction between the covariate and the scaffold treatment was not significant. The effect of the covariate was significant for claim and evidence, but it was not significant for the total explanation score or reasoning score.

³For both Mr. Kaplan and Ms. Hill, the effect of the covariate and the interaction between the covariate and the scaffold treatment were not significant.

⁴For Ms. Hill's students for evidence, the covariate was not significant and the interaction between the covariate and the scaffold treatment was not significant.

⁵For both Mr. Kaplan and Ms. Kittle, the effect of the covariate and the interaction between the covariate and the scaffold treatment were not significant.

⁶For both groups of Ms. Hill's students, the effect of the covariate and the interaction between the covariate and the scaffold treatment were not significant.

⁷For Ms. Hill's students for the multiple-choice scores, the covariate was significant, but the interaction between the covariate and the scaffold treatment was not significant.

^aLevels 1 and 2 are combined here because of space limitations. In the version used for coding, Levels 1 and 2 were separate for evidence and reasoning. For example, Level 1 included one piece of evidence and Level 2 included two pieces of evidence.