Here's something that really makes me mad.

Carroll, Violinguist, in the guise of a clinical psychologist, posted several anguished comments to this post, in which I describe the behavior of one of my not-so-good students. For the most part, Violinguist's remarks were addressed quite appropriately by other readers. I did not comment on them right away because: 1) I was gardening, and 2) they really pissed me off.

This is the remark that infuriated me:

Face it: in the grand scheme of things, your exam is not particularly important, nor is it a true measure of any student's aptitudes--particularly in statistics for aspirants to the social sciences (trust me; I know: I was horrendous in statistics in school but if I do say so myself I have blossomed into a darned good clinical psychologist).

From this statement alone it can be determined that the person writing the comment was neither (well) trained in the social sciences, nor a "darned good" clinical psychologist.

I fight the idea that statistics is irrelevant every day in my classroom. Not only is statistics important for social scientists, it is an important component of any liberal arts education. Understanding and dealing with variability in measurements is fundamental, critical, essential for anyone who wishes to be a rational contributing member of society. The media bombard us with statistics each and every day, on every news broadcast, in the papers, and on billboards. "Four out of five dentists recommend!" "The risk of Type II diabetes increases!" "Lifetime risk of breast cancer!"

Without an understanding of statistics, statements about health risks, public opinion, government, public policy, effectiveness of different laundry detergents (and on and on and on) must simply be taken as fact or rejected as lies. Arriving at a real understanding of how this information may or may not influence each of us personally requires statistics. To be able to sift good information from bad information requires statistics. To be a good citizen requires statistics. "In the grand scheme of things" statistics may be one of the most important classes a student ever takes.

Let's talk about voting, the most important responsibility a citizen of a democratic society exercises. Modern (and not-so-modern) technology has provided us with an array of different voting machines. Should we worry about how secure these machines are? When should we become concerned about vote fraud? Are our voices being heard? Should we have to rely on some expert to tell us that an election was fair or should we be able to figure it out for ourselves?

Of course we have to be able to figure it out for ourselves. This is the purpose of the exit poll. The exit poll is a sample of voters taken at the polling place immediately after they have cast their ballots. The actual vote count should be similar to the figures provided by the exit poll, within some margin of error. Exit poll statistics that are widely different from the actual vote count signal election fraud. The exit polls strongly suggest that the 2004 election that returned President Bush to office was flawed. If we were citizens of the Ukraine, or Rwanda, the United Nations might have declared our election results invalid and sent in observers to insure a fair election. Where is the outrage in our country? No outrage here: we are so poorly trained as consumers of information that we would prefer to watch NBC's Medium than the election returns.

On to the psychologists. When I teach graduate students the clinical psychology students are especially well-aware that statistics is a key component of their discipline. Not only do they need to understand statistical methods to conduct their research, but even those who do not intend to pursue a research career appreciate that without a good understanding of statistics they will be unable to read and understand the literature in their field. How would you like to go to a doctor who is so poorly trained that she must rely on drug companies or newspapers to tell her what drugs are more or less effective in treating a disease? The "darned good clinical psychologists" know statistics well and use it to make themselves better therapists.

I appreciate that not many people need to know how to do an analysis of variance, but I insist that everyone needs to know what a standard (margin of) error is and how a margin of error constrains the claims someone can make about their data. To say that statistics is "not particularly important" reflects a gross misunderstanding about how information is marketed and consumed in our society and a terrible misconception of the work of social scientists.

I went away, but now I'm back.

I took a few, several, lots of mental health days. I watched Star Trek: The Next Generation reruns. I gardened. (Why are my zucchinis dying on the vine?) I read. I knit. It was nice, and I'm not ready to be back in my office but I have to be so here I am.

This morning I had a wonderful dream. We had an emergency national election to recall Bush from office. Bush lost, and just like that, presto-chango, Kerry was president. It was amusing how similar the emotional overtones of the dream were to those of dreams in which I figure out how to fly.

And how is your summer going?

I had one student this past quarter who, because of a family emergency, could not take the final and so received an incomplete. It is probably not too surprising that this student was struggling in the class and would not have done well on the final even if exigent circumstances had not arisen. I find that the better students would rather take their exams and get them out of the way, early even, rather than leave their coursework unfinished, regardless of their personal disasters. I had a student last year who was being shipped overseas to the war, and he took his final three weeks early rather than take an incomplete.

My struggling student has now contacted me (via email) to take her final and resolve her incomplete. However, her inquiries into how best to resolve the matter are a little strange. Rather than, "When can I take my final?" she asks, "When do I have to have my incomplete resolved by?" Refusing to play into this (guidelines about incompletes are easily found online), I instead responded (cheerily), "Come in and take your final any time! Are you free this afternoon?"

Something tells me she won't be in touch for a while. Should I be astonished that she has not spent the past two months studying for her final? Or that she's in no rush to start studying now?

Time for a change.

Reading over my more recent blog entries, I see that my summer has taken a nosedive into... well, I don't exactly know into what. So I wanted to write today about something good.

I have a student working with me for the summer. He is a rising sophomore, in the Honors program, and almost delightful to have around. I say almost delightful because, in the course of our day-to-day interactions, I have learned that he is completely incapable of perceiving when I am busy and when I am not. He will hover in my office doorway for extended periods of time, even when I am deep in discussion with someone wearing shiny shoes and carrying a briefcase. Similarly, he is incapable of distinguishing between problems I am interested in (e.g., "If we implement a model of the process in this way, I can't fit the data any more") and problems that I am not ("The program stalls at about this point in the code and I don't know why"). Oftentimes the hovering is only to report that one of these uninteresting problems has been solved.

I overlook these annoyances because he's really and truly brilliant, one of the few really brilliant people I've ever met. Over the past 6 months or so he has taught himself parallel programming in Fortran 90 and how to use R. He has learned some really difficult stuff involving stochastic processes and Bayesian modeling and implemented the stuff in raw F90 code. (Note that he's only taken introductory-level statistics.) He programmed up a model and had it working (fitted to data) in about a month, after I myself had tried for six months to do exactly the same thing and failed.

Like most geniuses, he is pure geek. When he first showed up on my doorstep, I handed him an old but serviceable computer and a Fedora distro. By the end of the afternoon he had his own workstation connected to the lab LAN. I must note here that he didn't know a thing about Linux or Unix. I had to tell him how to create a file. "This is a terminal window," I said. "Type 'xemacs.'" Now when I go in the lab he has a million xemacs windows up, some containing little bits of code, some of them running R, others editing data.^*

Now if only he would stop locking himself out of the lab.

^*Of course I would never choose my friends by their choice of text processing application, but if you happen to use an emacsen, well, I'll just be that much more willing to buy you beer.

I can't believe it really happened.

It's probably hard for someone who hasn't served as a graduate student's advisor to understand how rare and horrible this event is: I failed a Ph.D. candidate on his oral examination yesterday. While the obvious blame for such a failure may seem to lie with the student, it does not. It is the advisor's responsibility to see that his/her students are adequately prepared for their examinations, and to advise them to wait when they are not ready. The oral examination should be a rubber stamp. When it is not, it says more about the faculty advisor than the student who fails.

While my judgment of failure was sufficient to insure that the student would be required to revise and resubmit his thesis for reapproval (he gets one more chance), in this particular case each and every person on his committee failed him, including his advisor. I am now carefully following the Graduate School's policy for extricating myself from this train wreck. I have no intention of sitting through a second defense like this.