Open Education 2016

The low completion and certification rates in MOOCs lead some to question the learning effectiveness of these new learning environments. However, current reports show that the number of people who sign up for these courses has risen from an estimated 16-18 million to over 35 million users in 2015 compared to the previous year. This growth is also evident in the number of MOOCs offered, the breadth of topics covered in these new learning settings, and the number of universities collaborating to offer MOOCs across the globe. This is indicative of the fact that users see value in these learning environments that may be hard to detect using a single measure of completion or certification. Consequently, our research team has begun to explore available Canvas MOOC data in order to discover latent patterns among its learners and course design features that can support new and effective forms of knowledge production and learning.

On March 1, 2016, Instructure, the creator of the Canvas learning management system and the Canvas Network MOOC platform, released a de-identified dataset from Canvas Network courses that were offered from January 2014 to September 2015. The data was queried, organized, and de-identified using a process similar to the one used for the HarvardX-MITx Person-Course data release of 2014, and then the data was made available to researchers on Harvard's Dataverse service. Instructure has opened access to this data to create opportunities for identifying and solving educational challenges in online learning. The dataset includes over 325,000 aggregate records, with each record representing one individual's activity in one of 238 Canvas Network courses. The variables available in this dataset include administrative variables from the Canvas Network system or computed by the research team, as well as variables generated by the users either through their interaction with the course or collected through surveys.

In line with the call for the need for "better data" to help understand the experience of MOOC stakeholders that extends beyond the number of clicks users make as they interact with a MOOC platforms, our analysis includes contextually rich data in the form of qualitative course metadata, course review notes by MOOC designers, and instructors' end of course feedback on their experience with Canvas Network.

After reviewing the data set from Canvas Network, a single course from that larger set of courses was selected and comparisons of this course v. the larger dataset will be shared. The #HumanMOOC was selected for this review as the authors have worked together on the course and can share their stories from a quality review and course facilitation perspective.

This research is exploratory in nature. Hence, no formal a priori hypotheses have been formulated or tested. However, by exploring data from multiple sources, we hope to be able to formulate explanatory relationships that can be examined in future research efforts. During the session, we'll be sharing our initial insights gleaned from our effort at making sense of these diverse data sets as well as next steps and futureplans.