This report documents the results of the investigation into the language learning potential of data-driven teaching materials on source use for undergraduates in a college-level writing course at a large land-grant Midwestern university. The investigation is a part of a large project which comprises three major stages: linguistic analyses on source use of 149 documented essays written by college students, development of data-driven materials on source use, and evaluation of the materials. The data-driven teaching materials consist of a corpus-based web tool and a computer-delivered online lesson on source use. The corpus-based web tool provides examples of citing sentences in the collection of 79 A-graded essays as concordance lines which help illustrate different features of source use, and displays graphs showing frequency distributions of citing sentences across sub-categories of each feature of source use. The computer-delivered online lesson contains two major tasks each of which has questions that guide students to observe the use of a feature of source use in the corpus-based web tool. This report summarizes key findings of the implementation of the materials in a naturalistic instructional setting. These findings focus on the language learning potential of the materials which concerns two major aspects: (1) whether the pedagogical design characteristics of the materials led to the students’ hypothesized learning processes (i.e., noticing and focusing on features of source use), and (2) whether the students gained any knowledge, skills, and awareness about source use after the training.