Tuesday, July 22, 2008

Want advice? Take stats.

That’s right. There are relatively few life-courses I can imagine that wouldn’t be made easier by a decent understanding of at least introductory statistics. A frankly, there are a lot of lives that would be improved greatly by an excellent understanding of more advanced methods. I would include mine in the latter category, fortunately or unfortunately, depending on your perspective.

If this summer has taught me nothing else, it’s that there are not enough hours in a lifetime to take all of the statistics and methods courses that one should. (I use one here rather generally, knowing that many people are uninterested in research. Still, I think it’s good for everyone to know, at the very least, what it means when the New York Times reports that something is “associated” with something else, rather than something “causes” something else.) That said, I’ve also discovered that even the most lucid and clearly written stats books lose something without a lecturer or a discussion with someone who already understands the stuff.

Additionally, if you can learn about 5 different stats packages, this is helpful. Everyone seems to use a different one, and I haven’t met anyone who could, with a straight face, say “No, I don’t mind at all if you work with my data using a statistical program I do not like.” So this summer, in addition to pioneering my way through large data sets of categorical data, I’m learning Stata, one of the friendlier stats programs I’ve encountered in my brief tenure as a stats program victim user.

Bottom line: take stats. Secondary take-home message: send me happy thoughts of correct Stata syntax and successful analysis.

No comments:

Post a Comment