Here is my final report on the Statistics One course that finished last week. After covering variance, correlation, and single/multiple regression in the first half of the course, we learnt about mediation, moderation, t-tests and ANOVA in the second half. The last three weeks were packed with new concepts for me and I felt we fairly galloped through some of the lectures to cover all the contents as scheduled.
We also continued analysing data in “R” (an open-source statistics package) and I started writing my scripts in RStudio, after seeing it recommended on the course discussion forum. RStudio brings together different workspaces in a single environment, providing a more user-friendly display than “R” itself.
So was it worth it?
Yes, definitely. The course had some glitches (bugs in quiz marks, typos in “R” scripts, lectures uploaded too late) and students were quick to criticise on the forum, but I’m not complaining. It was free. My investment in terms of time was about 50 hours over the last 6 weeks and as a result I’ve improved my understanding of the statistical methods and analyses that I translate in clinical trials. I may never have to set up regression models for a mediation analysis ever again, but my target language writing in statistics is going to sound more natural and have a degree of clarity that it didn’t have before.
Apart from polishing my statistics’ skills, I also took away a very positive experience in social media leverage of MOOCs (massive open online courses). Students contributed actively to the course forum, discussing assignments and sharing useful resources. When the Coursera site was down for several hours in September, some users quickly set up a Facebook group to provide access to data files. Here are some interesting figures about the course and its social media impact:
- The public Stats One site got more than seven thousand Facebook likes.
- The course itself started off with over 75,000 students. It would be interesting to have some stats on drop-out numbers and pass rates, but Coursera is suspiciously silent on this sort of feedback.
- The course forum had 100 pages of threads, some of which received more than 14,000 views and 500 posts.
- The Stats One Facebook page registered over 1,500 members in the first 48 hours.
- Google + didn’t get so far: only 46 people added the Stats One page to their circles and just 2 threads got going before it fizzled out.
- Twitter has been alive with #coursera tweets (especially when free online education was made illegal in Minnesota a couple of weeks ago).
- And, of course, I wasn’t the only person blogging about the course. Articles were factual, resourceful and critical.
In a nutshell, you can take away what you like from this course. It’s like stats itself, as summed up in the cartoon above and the multiple choice question below:
If you choose an answer to this question at random, what is the chance you will be correct?