New Research Instruments to Reflect Perspective-Taking and Student Academic Language Skills

• In Event: Integrating Research, Practice, and Design: A 10-Year Collaborative Effort to Improve Middle-Grades Reading Comprehension

Abstract

Efforts to test the novel hypothesis about contributors to reading comprehension required developing instruments to reflect student performance in the domains of academic language skill and perspective-taking. Academic language skill, to the extent it had been focused on in the measurement world, was assumed to be equivalent to vocabulary; a primary goal of the instrument to be developed was to tap aspects of academic language in addition to simple vocabulary: understanding markers of epistemic stance, discourse markers, the organization of longer discourses, register differences, and so on. A comprehensive assessment, the Core Academic Language Skills Instrument (CALS-I), was developed, shown to be sensitive to student age/grade, to have good reliability and internal consistency, and to be predictive of reading comprehension over and above a vocabulary test.

Previous measures of perspective-taking had primarily focused on the accuracy of student judgments about social events; the goal of the newly developed measure was to reflect student perspectival acts as they are actually used in the comprehension of text, as well as the deep understanding of the complexity of others’ intentions. Thus a multi-actor scenario-based assessment was developed, in which students were asked to take the perspective of different characters in the scenario and report what those various characters would likely do in that situation, why they would do it, and what the likely response would be. This assessment has also now been extensively validated, and been shown to be developmentally sensitive, distinctive from the CALS, and to predict reading comprehension.

Evidence from a series of research studies indicates that these two measures constitute important contributions to the field. First, they represent accessible ways to assess dimensions that have now been shown to relate to reading comprehension, and that should thus be incorporated in future studies attempting to explain variability in comprehension outcomes and the impact of interventions. Second, the availability of measures reflecting these skills will help create a new focus on them in instruction and the design of curricular resources.

Authors

• Paola Uccelli, Harvard University

• Robert L. Selman, Harvard University