College students rarely engage model-based epistemological reasoning about scientific data and evidence. The purpose of this study was to (1) investigate how scaffolding treatments influenced college students' epistemological reasoning about how data are used as evidence, (2) describe students' epistemological reasoning practice over the course of the study, (3) learn more about relationships among students' domain knowledge, epistemological beliefs about scientific knowledge, and epistemological reasoning, and (4) investigate how scaffolding for epistemological reasoning influences knowledge gain. Participants in this study consisted of three-hundred fifteen undergraduate students; all were juniors and seniors and all students were enrolled in one of two introductory genetics laboratory courses. Study participants included non-majors (Experiment 1, N =143) and majors (Experiment 2, N = 172). A partially mixed-methods sequential research design was used in this study; qualitative and quantitative phases were mixed during data analysis. A distributed scaffolding system was used in this study. All participants from each laboratory section were randomly assigned to one of three treatments; no scaffolds, domain-general scaffolds, or domain specific scaffolds. Study variables included domain knowledge, epistemological beliefs about the nature of scientific knowledge, and epistemological reasoning, scaffolding treatment was the manipulated variable. Findings were: (1) Chi square analysis indicated no statistically significant differences in epistemological reasoning by scaffolding treatment; model-based reasoning was not observed in students' explanations; (2) Spearman rho indicated no change in epistemological reasoning over the course of the study, however, statistical significance was not reached, however, a repeated measures ANOVA with Greenhouse-Geisser correction indicated a statistically significant within subjects change in epistemological reasoning, implications are discussed; (3) statistically significant bivariate correlations were found and (4) ANCOVA indicated pretest domain knowledge was a statistically significant covariate for posttest domain knowledge and a statistically significant main effect for scaffolding treatment was reached by Experiment 1 participants but not by Experiment 2 participants. Implications for instructional design and future research are discussed.