On the Law and Letters Blog, Belle Lettre goes into some detail on exactly how much she hates the statistics software program STATA. Below is an excerpt from her rant (slightly edited to preserve our pseudo PG rating):
STATA is the worst program in the history of the world, and if I didn’t need to know how to read/do statistics better, I’d want to drop the class. If I wasn’t concurrently enrolled in an awesome Empirical Analysis of Gender Discrimination course, I would definitely drop it, except that I actually want to get an article out on sexual harassment by the summer, and I sort of argue for the use of empirical analysis of discrimination by courts and the EEOC.
Also, while the professors and the TA are awesome, I do not really understand the pedagogy of belaboring the basic DOS commands (granted, I suck at those, but dude, just give a handout) and rushing through, in the last half hour of class, the section on how to tell Stata to calculate the fraction of observations that are one standard deviation away from the mean for the two variables it took me three hours to figure out how to generate and code as dummy variables. I cannot, for the life of me, recall learning how to do this in class. Nor how to properly do a histogram and specify heights and widths and $#@&. I think I can calculate this stuff with a calculator and draw boxes on graph paper faster, but I have to send in log files. Which also took me a half hour to figure out how to do.
Thus, I do not get the point of making people muddle through basic DOS-like commands without instructive guides. Isn’t the point to teach the statistics, not “learn how to be frustrated at a &*$% irritating program”?! Tell people what stupid monkey commands to enter, but ask them to analyze the significance, which is the important part, right?
First, I’d like to say that I really like her phrase “stupid monkey commands” - it seems like a phrase that people could use in everyday conversation - much like another phrase I just learned, “leave Britney alone!” which can be used against someone when they are giving you a hard time (e.g. your paper discussant on a conference panel). But I digress.
Belle, stupid monkey commands and feeling frustrated and inept for no good reason are all an integral part of the development of a social science scholar. It’s kind of like the Socratic method in law school - there’s not a lot of solid evidence that it actually makes you a good lawyer and it’s usually pedagogically inefficient and even abused by some profs. But, it establishes the hierarchy and makes you “work for it”. If everyone just gave you the STATA code or taught their law classes in a concise, straightforward manner, then you wouldn’t learn to “think like a social scientist (or lawyer).”
All fun aside, there is a “hide the ball” dynamic in almost every professional field, but hate the game, not the player (i.e. STATA). It’s funny that STATA should be criticized for being nonintuitive; when people came to discover it in the mid to late 1990s it was heralded as a great advancement, because it allowed you to perform some relatively high end statistical analysis (much more so than SPSS) and was much easier to use than other dos command based programs (e.g. LIMDEP, SAS, EVIEWS etc.). Now, STATA even has a windows based interface, although it’s not that great yet.
For what it’s worth there are some great books for the beginning STATA user - they’re kind of like the Emanuel’s law outlines of statistics. These can be found in the STATA bookstore, although you might find used copies cheaper on Amazon or elsewhere. I have found Lawrence Hamilton’s series of books very helpful. STATA also runs an email listserve where you can post your questions, but that can mean a lot of emails unless you take steps to get it in batches. I also find the UCLA statistics tutorial to be very helpful. Finally, and this is most important Belle, you need to convince your law school to hire more interdisciplinary legal scholars - not only will it help you with these STATA problems, but it will give bloggers endless fodder for debate.
Meanwhile, I think that this little video of “Nick Burns the Computer Guy” is relevant and appropriate to the topic.
3 responses so far ↓
abwhitford // January 21, 2008 at 6:33 pm
One letter: R.
prisonrodeo // January 21, 2008 at 7:29 pm
What Andy said.
[Tomorrow I'm teaching my students how to xyplot(internalwar~polity,panel=function(x,y){panel.xyplot(x,y);panel.loess(x,y,span=1);},xlab="POLITY",ylab="Internal War Indicator")].
bulldog20 // January 22, 2008 at 12:00 am
That guys has obviously never used minitab… he’d be cursing til the sun came down if he had.. Stata ain’t perfect, but it’s easier than most.
You must be logged in to post a comment.