This study extends previous research that explores how visualization affordances that computational tools provide and social network analyses that account for individual- and group-level dynamic processes can work in conjunction to improve learning outcomes. The study's main hypothesis is that when social network graphs are used in instruction, students receive otherwise hidden information that influences more strategic decision making about whom to interact with in order to gain more knowledge. This in turn can influence a shift in how students interpret the nature of socioscientific issues toward a more complex understanding. Results of the intervention show that among this population of 76 Grade 7 students, rules by which students selected whom to talk to in paired discussions about a complex socioscientific issue shifted from nonreflective or socially driven mechanisms (e.g., waiting for any random person) to reflective or information-driven mechanisms (e.g., specifically choosing someone who had a lot of knowledge). Results are compared to research in academic domains other than education that supports their robustness as decision-making strategies. (Contains 4 figures, 1 footnote, and 7 tables.)