Research on multiple text comprehension reveals key principles and elements of comprehension: readers' mental representation, cognitive text processing, and strategy use while reading multiple texts (Goldman, 2004; Rouet, 2006). However, many studies of multiple text comprehension fail to investigate the influence of reader bias. Grounded in both the literature on reading strategies and the social psychology literature on bias (e.g., Edwards & Smith, 1996), this study investigated how readers' topic beliefs influence comprehension strategies in relation to bias.

The participants for this study were 15 undergraduate students, chosen as they represented three distinct topic beliefs related to the Israeli-Palestinian conflict. There were 5 pro-Israel, 5, pro-Palestine, and 5 neutral participants. While thinking-aloud, participants read two maps and five texts about Israeli settlements in the West Bank, and the ongoing Palestinian-Israeli conflict. The texts and maps were presented in the iMTC (internet-embedded Multiple-Text Comprehension measurement tool) environment (Kim & Cho, 2011). In addition, measures of participants' prior knowledge and topic beliefs were gathered, while their reading times and Internet searches were recorded by the iMTC. Participants' verbal reports were coded based on existing coding schemes for reading strategies (Goldman et al., 2012; Pressley & Afflerbach, 1995). Five families of strategy were determined: Considering text content, Acceptance and resistance, Monitoring, Evaluation, and Information need and search.

The study has three major findings. First, initial belief differences between groups of different beliefs increased after reading, meaning that participants showed biased assimilation processing during reading. Second, the participants' biased processing was not detected in the three types of reading measures: reading times, reading orders, and Internet searches. Finally, the study found that participants with different topic beliefs showed different strategic patterns in relation to bias. In particular, acceptance and resistance distinguished the three participant groups' strategic processing. Participants accepted belief-consistent text information and resisted belief-inconsistent text information. In addition, three cases of participants' biased strategy use were qualitatively analyzed. The analyses demonstrated that participants' topic beliefs played a role in creating an interpretive framework that evaluated, accepted, or resisted information during reading. The findings, limitations, implications for future research and instructional practices are discussed.