Predicting high school students' argumentation skill using information literacy and trace data

Strong information literacy, collaboration, and argumentation skill are essential to success in problem-based learning (PBL). Computer-based scaffolding can help students enhance these skills during PBL. In this study, we investigated predictors of the quality of arguments high school environmental science students wrote in support of their solutions to a PBL problem. Specifically, we used Bayesian regression to examine how information literacy, collaboration, and time spent and number of words written in various sections of scaffolding combine to predict argumentation quality. Significant positive predictors of argument quality were information literacy posttest scores, individual work time, and number of words typed in response to prompts in the information literacy section of the scaffold. Significant negative predictors were group work time, number of words typed in response to prompts in the 'define the problem' section of the scaffold, and time spent in the 'define the problem' and 'building arguments' sections of the scaffold.