samedi 31 août 2013

Data Scientist Intro Education, 6,800 at a Time

A couple of months ago I wrote about Dr. Lisa Dierker at Wesleyan University in Connecticut and her plans to teach statistics to 13,000 students from around the world in a free massive open online course (MOOC). A professor of psychology at the university and chair of the Quantitative Analysis Center, she designed a course in which students could choose projects on topics that interest them.


Early this week I checked back with her to see how the course had done. The university supported her by paying several students to help mentor online participants, which turned out to be a great experience for both. Student monitored the discussing forums for large blocks of time during the week, responding to posts and launching conversations.


“Wonderful things can happen when the instructor doesn’t jump in,” she said. She often would let the conversation go and several hours later weigh in and have everybody move on.


“I think it was fantastic,” she said. “Our students who come back to mentor and tutor in on campus intro courses learn more than the students in the course. It’s not just that this is the second time around for them. They get the information again, but now they have to use it flexibly, they have to be able to respond to all the questions the students are asking. In any technical area we often make students take an intro course and then they have no interest in going on. This was a great opportunity to think how peer tutoring can be that second course — getting students exited, getting their confidence up.”


It was also a great exercise for her, she added.


“When you have to think about 10,000 students you will never talk to or explain things to, it took my materials to a new level. This had to work without my being able to talk to them.” She took a project approach to the MOOC, she added, with videos, lectures and demonstrations so students could go back to any areas they had difficulty with and collaborate with other students online.


The course drew a wide variety of students. A politician took it so he could better understand policy, and a nurse took it so she could go toe to toe with hospital administrators and researchers. A math Ph.D. from Spain signed up because he was teach a lot of statistics courses to earth and environmental science students, and he wanted to make his own course more interesting.


“He ended up being a super TA, helping everyone. A SAS programmer from Chicago who has a job in data management and then pumps his results to Ph.D. statisticians to do cool analyses, took it so he could learn how to use the data himself. After I showed students how to do something, he would show five other ways that I had never seen before. I have never in my life taught a course where the actual enrolled students brought so much to the table.”


The discussion rooms were super active, she added, although she saw the same names appear again and again, much like a regular classroom.


“I bet only 300 were on those discussion forums. They can be an amazing learning tool for those engaged, but just as a classroom or office hours, if they don’t walk through the door those things are not useful.”


The analytics company SAS made its software available online for the course.

“They were amazing at supporting it all around the globe,” said Dierker, “making sure the data was up on their cloud servers, but there were people who didn’t have strong enough computers, especially outside the US.”


Ron Statt, a senior director in R&D at SAS whose team supported the project and worked closely with the professor, said the course offered a great learning experience. SAS carried on a collaboration with the professor throughout the course and participated in the discussion forums to offer support, share advice, and get a pulse for what was going on in the course, which not so incidentally introduced thousands of students to SAS software.


“Students came from all over the world, but we saw the largest participation from the United States, India, Britain, and Canada,” said Statt. “We saw students discussing issues such as not having appropriate systems or wanting to use their Macs. It so happens we had a solution for them. To make it easier to program in SAS, last year we introduced SAS Web Editor, a Web-based tool for writing and running SAS code. Available at no cost, SAS Web Editor requires no software installation. Users simply connect to a website to access code, files, projects and libraries, anytime and anywhere. Since it only requires a Web browser, it works on Macs as well as iPad 2 and above. We let them know about SAS Web Editor and many students began to use it. It resolved their issues and enabled them to move forward with the MOOC course.”


Although MOOCS have been criticized for high dropout rates, Dierker said 6,800 completed the course. The most common reason for dropping out was simply a lack of time; most of the students were working or going to university.


Back in December when colleagues knew she was going to teach the course, she was inundated with encouraging articles about what a hare-brained disaster MOOCs were. She got a little concerned about her decision but now has no regrets, although she exhibits some becoming humility about the timing.


Hers was a first generation MOOC experience, she said.


MOOCs could change everything


“I did it at a great time,” she said. “This is exploding. It was a first generation MOOC experience, but MOOCs could change everything. We are getting a better rap than we might deserve because we had such willing participants,” she added. “I am taking the positive feelings with a grain of salt. The students really wanted to do this, and although the course was designed for complete novices, a lot of the students were data geeks or wannabe geeks.”


She expects the next generation MOOCs will have ways of requiring participation from students and eventually ways to test them and offer credit.


MOOCs are raising big questions about how education is done, she said. One colleague at Wesleyan suggested the world might not need more than two or three intro courses in statistics or biology, a suggestion which did not make him the most popular man on campus. Her experience designing a course for delivery online has made her think about the usual hierarchy of introductory courses followed by more in-depth courses. The modern world is too dynamic for a simple ladder of courses; some professors are talking of staged courses where students can come in and pick from the materials they find most relevant to their needs. Now she is excited about the idea of marrying a MOOC, on-campus course and peer-tutoring in ways that feed each other rather than have three separate courses to support.


SAS’s Statt said that discussing the course with Dierker before, during, and afterwards enabled SAS to anticipate issues, provide support in near real-time, and discuss what we learned and how we might make future enhancements.







via Business 2 Community http://www.business2community.com/tech-gadgets/data-scientist-intro-education-6800-time-0597862?utm_source=rss&utm_medium=rss&utm_campaign=data-scientist-intro-education-6800-time

Aucun commentaire:

Enregistrer un commentaire