Argument Evaluation, Reading Strategies, and Opinion Change in the Digital Environment

• In Event: Reading in the Digital Age: Affordances and Consequences for Learning, Judgment, and Performance

Abstract

Purpose/Objectives
The present study examined Twitter’s priming effects on argument evaluation, reading strategies, and opinion change.
Theoretical Framework
Social Media Technologies (SMTs) are web-based and mobile applications that allow individuals to create, engage, and share content in digital environments (Leonardi, Huysman, & Steinfield, 2013). Twitter is a popular SMT and a powerful medium of opinion formation and change (Howard, Duffy, Freelon, Hussain, Mari & Mazaid, 2011). In this study, we ask if opinion formation and change is primarily due to the medium itself or the content of the message.
If the medium itself does have power over the content, then approaching the medium, regardless of the content, will prime specific cognitive processing, much like different text genres do. Indeed, genres influence the processes and strategies readers employ during reading (Zwaan, 1994). If SMTs function as genres, then they likely exert their influence by priming specific reading processes and strategies.
Method
Two experiments were conducted. In Experiment 1, 75 undergraduate students were asked to read and evaluate a series of 40 arguments on the Scottish Independence Referendum, a topic they had limited prior knowledge on (10 strong “yes”, 10 weak “yes”, 10 strong “no”, 10 weak “no”). Half of the participants read the arguments framed as tweets, while the other half read them as plain text. Participants also were asked to state their position (for or against) before and after reading the arguments. Further, participants were asked to identify the reading strategies (strategies scale; Bra˚ten & Anmarkrud, 2013) they used during reading, as well as their usage of Twitter. In Experiment 2, 74 participants completed the same procedures as in Experiment 1 with one modification: the arguments were revised to refer to the Scottish Union (instead of Scottish Independence) to examine a potential ‘negativity bias’ finding in opinion change obtained in Experiment 1. Both experiments were conducted online using Qualtrics.
Results
In Experiment 1, media priming did not influence argument rating, F (1, 72) = 2.94, p = .091 or strategy use F (10, 60) = 1.64, p = .117. These results were replicated in Experiment 2, suggesting that Twitter did not prime certain processes or strategies when compared to plain text.
The analysis also showed before-after opinion change, χ2 = 16.11, p < .001. Specifically, 17 out of the 45 “yes” (37.8%) in Experiment 1 changed to “no”, and 4 out of 28 “no” (14.3%) changed to “yes”; 52 (71.23%) did not change. These results suggest a ‘negativity bias’, namely a tendency to assign greater weight to negative entities (Rozin & Royzman, 2001). The reframing of arguments in Experiment 2 confirmed this finding, as it reversed the opinion change pattern, χ2 = 8.64, p < .003. Specifically, 8 of 29 “yes” (27.6%) changed to “no”, and 12 out of 34 “no” (35.3%) changed to “yes”; 43 (68.25%) did not change.
Significance
The findings suggest that merely being primed to SMT is not sufficient to influence argument evaluation, reading strategies, or opinion change. Ultimately, negativity bias was the predicting factor of opinion change.

Authors

• Bader IMNM Mohsen, University of Minnesota - Twin Cities

• Panayiota Kendeou, University of Minnesota