| A number of universities offer fully web-based language courses. Many others are considering this option. Yet it remains unclear whether web-based courses can be as comprehensively effective as classroom-based courses. And, if so, what might an effective web-based language course look like? This dissertation considers the German Online at PSU program, a fully web-based basic language program (first through third semesters) that was introduced at the Pennsylvania State University in 2006. The first half of this dissertation outlines the program and its empirical underpinnings. The second half consists of a comparative, quasi-experimental study of developmental outcomes and learner characteristics. All participants (N=33) were enrolled in a 15-week German course. Each week, classroom-based learners met with their instructors and classmates for four, fifty-minute sessions. These sessions included group study of text and audio, speaking activities, and dyadic and small group discussions. Web-based learners never met with their instructors or classmates face-to-face. Instead, they recorded weekly speaking assignments and completed two sixty-minute, text-based chats each week. All other aspects of the courses (e.g., automated grammar exercises and mobile immersion activities) were the same across the two conditions. Learner development was measured by a variety of pre- and post-tests, ranging from standardized assessment measures, such as the WebCAPE and the SOPI (rated according to a 50-point scale), to more experimental tasks typical of laboratory-based studies in cognitive processing (speeded translation recognition and speeded grammaticality judgment). Correlations between individual learner variables (age, SAT scores, semester standing, previous course grades, and phonological working memory) and development were also considered. On all measures, it was found that web-based and classroom-based instruction supported statistically significant and comparable development. Some correlations between development and learner characteristics were found, but no correlations were found in both conditions at the same time, and the lack of repeated correlations between particular individual variables and all developmental outcomes suggests that no characteristic single-handedly determines the level of learner success in either environment. This dissertation provides quantitative evidence that web-based courses can constitute a viable and comprehensive alternative to classroom-based language instruction. |
