MODELING MULTIPLE SOURCE USE: USING LEARNER CHARACTERISTICS AND SOURCE USE BEHAVIORS TO PREDICT RESPONSE QUALITY
Alexander, Patricia A
Multiple source use (MSU) has been identified as both a critical competency and a key challenge for today's students, living in the digital age (Goldman & Scardamalia, 2013b). Theoretical models of multiple source use provide insights into how the MSU process unfolds and identify points at which students may encounter challenges (i.e., in source selection, processing, and evaluation, Rouet & Britt). However, understandings of MSU have been limited by two gaps in the literature. First, while points of challenge in students' MSU process have been examined independently, comprehensive models considering the joint role of source selection, processing, and evaluation in task performance have not been fully investigated. Further, while research on MSU has focused on students' behaviors when engaging with texts, individual difference factors have been considered only to a limited extent, despite their theorized importance (Rouet, 2006). The purpose of the present study was to examine the extent to which multiple source use behaviors (i.e., source selection, processing, and evaluation) and learner characteristics (i.e., prior knowledge, domain general source evaluation behaviors, stances on the target issue) predicted open-ended task performance, both independently and in conjunction with one another. Participants were 197 undergraduate students, asked to complete measures assessing their prior knowledge, stances on the Arab Spring in Egypt, the topic of the task, and domain general source evaluation behaviors. Then, participants were tasked with using a library of six sources to respond to a controversial prompt about a contemporary event (i.e., Arab Spring in Egypt). While students engaged with sources, log data of source use were collected (e.g., number of sources accessed, time on texts) and participants were asked to rate sources accessed in terms of trustworthiness, usefulness, and interestingness. Four indices were used to assess open-ended response quality: (a) word count, (b) the number of arguments included in students' responses, (c) scores on the SOLO taxonomy (Biggs & Collis, 1982), reflecting the extent to which students' responses integrated and evaluated information presented across texts, and (d) the number of citations in students' answers. Key findings included the role of students' ratings of source interestingness and time on texts as predictive of open-ended task performance. Further, students' accessing of document information about sources (e.g., author credentials that may aid in source evaluation, Britt & Aglinskas, 2002) and trustworthiness evaluations were found to be associated with SOLO scores. Overall, as compared to multiple source use behaviors, learner characteristics were found to have a more limited effect on task performance. Findings are discussed and implications for theoretical conceptions of multiple source use and instructional practice are presented.