Course: I Heart Stats
School: Notre Dame (via edX, free)
Instructors: Dan Myers
Quote:Statistics can be confusing and opaque. Symbols, Greek letters, very large and very small numbers, and how to interpret all of this can leave to feeling cold and disengaged—even fearful and resentful.
But in the modern information age, having a healthy relationship with statistics can make life a whole lot easier…. The purpose of this course, then is to help you develop a functional, satisfying, and useful life-long relationship with statistics….
What you’ll learn:• Select appropriate statistical tests for data according to the levels of measurement
• Perform basic calculations to determine statistical significance
• Use standard methods of representation to summarize data
• Interpret and assess the credibility of basic statistics
I began to realize, midway through Genetics & Evolution, that I needed a much better understanding of statistics. Think of all the jokes you’ve heard: there’s no better way to obfuscate an issue than to come up with a statistic that sounds impressive, even though, when examined more closely, it doesn’t hold up, because few people know how to examine stats more closely.
So I took this course. Two MOOC-friends of mine took this course at the same time I did; they both loved it. For that matter, it seems a thousand people Liked the Facebook page. Me, not so much, but I say it every time I do one of these: every course I hate, someone else loves. In this case, a lot of people.
So what was my problem? Mostly, I just really, really hate stats.
In addition to arachnophobia (which actually contributed to my dropping the Animal Behavior course this week, believe it or not), I’ve discovered I have sigmaphobia: a fear of summation signs. Those are the Greek capital-Sigmas, the things that Greek restaurants use as capital E’s even though they’re really S’s. There’s a great line in Sylvia Plath’s The Bell Jar about protagonist Esther Greenwood’s visceral reaction to reading German: “…each time I picked up a German dictionary or a German book, the very sight of those dense, black, barbed-wire letters made my mind shut like a clam.” That’s how I am with summation signs. It’s just a personal quirk, and it’s necessary that I keep calm and carry on, but it makes things like stats extra-special difficult.
I hated all the calculating – and this course was about 85% calculating. Sure, adding a column of numbers, calculating an average, squaring each number in the column, summing the squares, etc etc, isn’t that hard – that’s what Excel is made for. It’s just incredibly tedious and I hate it. But… I do need to understand stats better, and I had to start somewhere.
I did learn a few things beyond calculating. I even recognized a few things I’d seen in other courses: “Oh, so that’s what he was talking about when he said he was using a stricter standard of significance testing… oh, that thing we did back then, that must’ve been a chi square.…” Just yesterday, I was watching someone talk about his research and he mentioned within and between factors, ANOVA and t-tests; I was so excited – I know what that is! I’d have to listen way more carefully to it to say I understood his research design, but at least I recognize the tests he’s using on his data, and that’s more than I could’ve said a month ago.
As for more comprehensive understanding, Dan made reference several times to more advanced courses for theory, and I’m going to need that. There was some attention to concepts in this class: discussion of data types and the requirements of the various stats, and a very good “Hypotheses Testing Q&A” with Dan and Sara early on (my favorite thing in the course), and the first half of the final was relatively conceptual: Here’s the situation, what stat should you use. But one eight-week intro course isn’t anywhere near enough, at least for a mathematical idiot like me.
Despite my antipathy towards the subject itself, the course did a lot of things very right. The calculations for each of the statistical methods – Standard Deviation, Chi Square, T-test, ANOVA, Regression, Correlation – were demonstrated three times by three different people in different formats, with PDFs of detailed Notes available as well, so if one route of explanation didn’t appeal, there were alternatives (my personal preference was for Sara’s runthroughs). Plenty of practice problems were available, both as practice and in graded form. The course even offers “I Heart Stats” t-shirts for sale, just like the one Sara wears in the videos. How many courses offer that? I’ve said before MOOCs should offer swag. I would’ve bought a bunch of course-specific t-shirts or coffee mugs or tote bags. Just… well, not for this course.
They did their job. And I did mine, even when it hurt. I’m kind of proud of that, that I kept going with a course I hated, since I’ve been dropping courses all over the place at the first sign of “This isn’t anywhere near as interesting as I thought it would be.” I finished. Granted, I kept an eye on the progress meter, and I stopped as soon as I’d done enough of the final for a pass (ok, not all that proud of myself), because I really didn’t want to calculate another σβ.
One of the minor quibbles I had was partly of my own doing, I think. I seem to recall being asked if I’d be willing to participate in a research study during the course. I love MOOCs so I always agree to this stuff – it’s possible it was another course, in fact, not this one at all. But throughout, little questionnaires kept coming up. “Checking In”: “Which emotions best describe how you’re feeling about the course?” with a list of maybe a dozen emotion words (anger, contentment, hope, isolation, shame). Then there was SAM, the Self-Assessment Manikin. It seemed like overkill to have two kinds of mood-assessment, but maybe one was the study I’d signed up for, or maybe they were thinking of people who weren’t fluent in English, or maybe they wanted to compare results between the two different representations (which is an interesting idea, by the way). Thing is, these things cropped up so often early on, I started to get really pissed off whenever I saw them… so I kept entering things like “angry” and “anxious” and “confused” because that’s how I was feeling, even though I was doing fine with the course material. Which is also kind of an interesting result. The instructors are sociologists, after all.
I also took great exception to one of the examples used, with self-described fake data showing differences between IQ scores for different races, with a reference to Murray’s book The Bell Curve, which to me was very controversial back in the day. I was pretty upset about this. It seemed, at best, insensitive to start flinging around fake data showing white people have higher IQ scores than black or Hispanic people. I was relieved when we went back to things like evaluation of popcorn brands, and relating hair products to gender or exercise to work productivity. I had some discomfort with how some results were phrased as well; I wish someone had said, at some point, “Correlation is not causation.” Because that’s where the fun starts. Maybe I just don’t understand the concept.
But overall, this was a detailed and effective basic introduction to a topic that befuddles a lot of us; if you want to know how to calculate ANOVA or standard deviations, strap on your calculator, crank up your spreadsheet, and go for it. Who knows – you might end up hearting stats.