The neverending statistics saga
For a frustrating and historical reason, our students are not necessarily required to take statistics. Even if we manage to convince them that they must do it for their professional development, they don't always take the class when they most need it. Which leads to the following situation.
Two students in my graduate class have finished their data collection, and need to analyze it. But they haven't taken stats, which leaves me wondering what the hell I'm really supposed to do. Their design is fairly simple, but even so, how are they supposed to understand the results when they don't know what an ANOVA is, much less an interaction or a planned comparison?
What really annoys me about this situation is that if any analysis is going to get done, I have to do it for them. That really pisses me off, because I begged the department to let the students rearrange their schedules so they could take stats, and everyone blew me off. So now I'm being punished. I'm meeting with the students in half an hour, at which time I will run their ANOVA, and then rearrange the data to do the comparisons, because SPSS is such a stupid program. And then I will explain to them what an interaction is--in their case, up to a 3-way interaction, which is not the easiest thing to interpret.
And as if this isn't bad enough, then they have to write up the results! I realize that's something I have to teach them how to do as part of my class, but even if I give them the results section from a few example papers, they're going to be mystified because they probably won't even understand what the authors are talking about.
Well, this situation just sucks, but at least I can probably use it as ammunition to really show the rest of the faculty WHY statistics is important for these kids, and why it must be a requirement for students in a certain subfield.
technorati tag: teaching-carnival
posted by Ianqui at 1:00 PM
6 Comments:
I don't know what your disicpline is, but it is hard to imagine a grad program in which students collect data but are not required to take stats. Amazing. And what a lot of extra work for you!
And yes, SPSS is a stupid program.
A current grad student of one of my advisors (the G one, not the B one) emailed me because she wanted to know if she should take the second half of the stats class (it's a two semester class). Apparently she had been told to ask me. I told her that if she was going to be doing experiments (and she is) there's no question that she should take it. I am a bit baffled by those who think that learning very introductory stats (probability theory , basic t-tests, and ANOVAs) is sufficient when you're *training* to do experimental work. I'm not sure why you'd want to stop at the introductory stuff when they offer the advanced stats for experimental design, unless it was absolutely clear that you're not going to do any further experimental work.
I just wish that her advisor, who is starting to do experimental work of her own would 1) take a class herself, and 2) educate herself about what the stats classes are teaching so that she would know if her students need it or not.
I hope you manage to get your department to make adjustments for your sake. Otherwise, it really seems to be hurting the students. If I was them, I would hate for my professor to do the stats and explain to me what they meant, because I don't think I'd really get it without a broader background in stats.
So what happened, you couldn't make stats a prerequisite for graduate students taking your class?
What the hell? Wow, when you started your post I thought you were referring to undergrads, which I could understand because while we make ours take "stats", the classes they take go out of their way to shield them from any actual math.
But grad students writing up data analysis on techniques they don't even understand? Do they have trouble finding faculty positions?
Yeah, it's sort of hard to explain without giving too much away, but until another prof and I got here (around the same time), there wasn't that much need for statistics. But this is my third year, and although I immediately began talking about this issue upon arrival, no one has really listened to me. The problem is that now students are either taking my class or doing research with me, and statistics is essential.
It's not really the students' fault--upon realizing what my kind of research entails, they were all totally willing to take stats, and some of them went ahead and took a class even though it pissed a lot of other profs off. The problem is that the rest of the dept isn't willing to budge re: their requirements. So profs were offering classes that are only offered every 2 or 3 years, and expected students to take it, when they also needed to take stats for their research. All of this has led to a giant mess.
Way to fight the good stats-dork fight ;).
My field has a similar problem (re: no/minimal stats) and I am a victim of it. I know what is possible to do statistically but actually doing it can put me in a bind. Apparently at the university where I am now working there is a really solid stats sequence which I am thinking of taking just so that I'm up to speed. How weird would it be to take stats with your own doc students (esp. if you are younger than your doc students). The stranger thing is half my field requires that you be fluent in awful things like differential equations and physics and require them as prereqs for everything and half my field doesn't care about any stats at all. Weird huh?
Post a Comment
<< Home