OpenLearn 2007 - OLI: accelerated learning with @ CMU's OLI (day 2)

Candice Thille and Joel Smith from Carnegie Mellon's Open Learning Initiative, explained their very interesting findings on the OLI's impact on learning. They measured the effectiveness of OLI statistics course in accelerating learning, and interestingly: students learn faster and better! And they are able to articulate in their answers, not just calculate them in statistical terms. I have discussed OLI more from a sustainability perspective here, but the approach of impact on learning is very interesting. And the findings even more so! Very promising.

A primary goal of the Open Learning Initiative (OLI) at Carnegie Mellon is to provide free access to high quality post-secondary courses (i.e., similar to those taught at Carnegie Mellon). Previous evaluations of the effectiveness of OLI courses have shown that our online courses teach students as effectively as existing instructor-led courses. Two such studies have found this result for the OLI-Statistics course. This report describes our current study of OLI-Statistics in which we are evaluating the accelerated learning hypothesis – that learners can learn a semester’s material in half the time, while still achieving the same or better learning outcomes.

The courses offered by CMU OLI are, as I have explained before, not ordinary open courseware materials. They mention the following about the learning material:

• based on learning theories
• use driven design
• transforming instruction, not transposing it
• scaffolded
• predict: immediate feedback through AI systems and cognitive tutors
• (expensive to make)

There are four feedback loops based on student learning data; science of learning, instructor activities, course design, student performance.. The initial assumption that the open courses would not be used at university appeared to be wrong: they were better than the normal courses, and quickly became a prototype to be used. The interesting thing is the research done on the learning outcome of students following the normal course, and students following the online course.

• Statistical literacy: The first research was set up between two groups of students: one group did the online version, and teachers were not allowed to guide them, only answering their questions once a week. The second group did the classroom version, and interacted in a classroom setting once or twice a week. One of the first outcomes was that there was not a statistical difference between the groups online and classroom learners.... but the application of the knowledge in real settings, or answering more applied questions was done better by the students doing the online version.
• Better and faster: The second research was done with a slight but important difference: the online students were allowed to interact in-depth with teachers twice a week. They would normally have prepared the questions, and had one or two hours of interacting with the teacher. Regarding the experience, both teacher (great! quality time with students!) and student (fantastic! I am learning!) were very contended. Concerning the outcomes, even better results were presented: the students were not only able to do it twice as fast, but their results on tests were significantly better as well.