Written by: Spencer Greenhalgh
Primary Source: Spencer Greenhalgh
This episode of Spencer Writes in the Library took place Thursday, March 26th around 9:00am.
Where am I working today?
I got to the library early enough that I could do something I’ve been meaning to for a while now: Claim a study group room! There are some scattered throughout the west wing of the library, but I’m currently set up on the second floor. As the name implies, the rooms are really intended for small groups of people, but if there are no small groups of people around to take it, there’s no issue with a lone graduate student stepping in.
What’s a perk of this spot?
It is nice to have a big but private space to yourself; plus, there was a chalkboard that I made use of at one point to sketch out some of my thinking.
What’s a problem with this spot?
Well, it feels a little too big for one person. I’ve been meaning to try one of these out for weeks, but I’m kind of disappointed. I’ll probably stick to desks and tables in the future.
What have I learned in this spot?
On my way up to the second floor, I got to take a glimpse at this set of “floral conversation cards,” which was pretty interesting.
How would I rate this spot?
3 out of 5 dentists. (Why dentists?)
What am I working on today?
More statistics work. I met with a consultant again on Monday, which was a good experience, even if there was some frustration involved. One of the things I’m exploring in grad school is how we can use existing data to answer relevant questions. That is, instead of designing and distributing an instrument in order to learn something about educational games, what can we learn about them from websites like BoardGameGeek, where an active community of learners has already collected a lot of information for us?
I’m convinced that there’s a great deal of value in this approach, but I’m learning that it makes statisticians (or at least one of them) uncomfortable. There’s simply too much we don’t know about how this data was collected, and if you want to rigorously meet the assumptions and demands of statistics, you need answers that these websites can’t provide.
At this point in my research, I’ve accepted that, and I’m figuring out what I can do that might be “close enough.” Today was about just that: Trying a couple of stats techniques that might work to see if we can at least come up with a reasonable approach to this study.
What’s the highlight from today’s work?
I got to run a principal components analysis today, and despite the difficulty I’ve had understanding this concept in the past, the results are actually quite cool. In short, this technique can be used for data reduction, or reducing a large group of variables to a smaller group of them. The interesting thing, though, is that the new group of variables is developed mathematically, and it’s up to you, as a researcher, to figure out just what they represent. I spent a few minutes today looking at the relationship of these new variables to the old ones to try to puzzle out what they were representing, which was a very interesting intellectual activity.