Using performance information obtained for program assessment purposes, this quantitative study reports the relationship of ePortfolio trait and holistic scores to specific academic achievement measures for first-year undergraduate students. Attention is given to three evidential categories: consensus and consistency evidence related to reliability/precision; convergent evidence related to validity; and score difference and predictive evidence related to fairness. Interpretative challenges of ePortfolio-based assessments are identified in terms of consistency, convergent, and predictive evidence. Benefits of these assessments include the absence of statistically significant differences in ePortfolio scores for race/ethnicity sub-groups. Discussion emphasizes the need for principled design and contextual information as prerequisite to score interpretation and use. Instrumental value of the study suggests that next-generation ePortfolio-based research must be alert to sample size, design standards, replication issues, measurement of fairness, and reporting transparency.
Keywords: ePortfolios, fairness, program assessment, reliability, validity