"Thou shalt not use thy knowledge of statistics to inflict harm"...oops.
I’ve been pretty stressed out lately over work, classes, other assignments, and...my statistics lab. As a result of this, I have also been very, very tired. Last night, I was staring at my computer screen, trying to generate intelligent answers to the lab questions and forcibly removing the cat from my desk every three minutes. I decided I needed a break and called IrishTenor to tell him my woes. After all, a boyfriend is supposed to lend a sympathetic ear when I want to complain about things that do not interest him and make his head hurt.
IrishTenor picked up the phone, and I started to tell him how sleepy I was and about my statistics lab.
IrishTenor: "Wait, is this something where I have to care?...Okay."
Me (reading from my assignment sheet): "Write the two-level hierarchical linear model for the main effects..."
IrishTenor (sounding dismayed and borderline violated): "And now you’re reading it to me. I was reading The Count of Monte Cristo."
Me (trying to conceal my amusement): "Should I let you go back to that?"
IrishTenor: "Yes. Goodnight, my dearest."
I went back to my computer. I saw instant messages from my ridiculously smart brother, who wanted to talk about psychometrics, statistics, and research methodology. Inwardly I groaned.
Those of you who know me know that these things are my main focus in graduate school and may be wondering why the prospect of discussing them with TheGoodOne made my head hurt. In fact, those of you who know me are probably wondering why I wasn’t literally jumping for joy. The thing to understand is that at any given time, 95-98% of my brain is devoted solely to keeping me from falling asleep in my chair. Also, TheGoodOne is a full-fledged, bona fide brainiac. I usually need to borrow about 50 IQ points from a neighbor or from someone on the street just to hold my own talking with him about anything. And we have completely different philosophies about approaching research, which methods to employ, how to employ them, and what kinds of questions we can ask.
TheGoodOne had many rather vague questions about coursework in research methodology, survey methods, psychometrics, and statistics. Fatigued and unable to clear the SAS syntax I was seeing inside my head, I finally asked him, “What specifically do you want to be able to do, with statistics?”
"Well, by the end of my political science stats training, I should have mastered regression in its various forms. But I'm looking for analysis other than regression."
There it was. My chance for small scale revenge. Because I am well enveloped in the nerdy evil that is statistics and because I like it that way, I felt enlightened rather than disillusioned when I finally learned this. I finally had a global understanding of all the things I'd been calculating and interpreting for years.
For my brother, I think this felt more like The Firm, when all the partners pull an associate into the office or conference room and reveal that they are merely a front for a Mafia money laundering scheme. I knew it would feel that way for him. But God help me, I said it anyway.
"Other than regression? Like what? Everything is regression."
No response. I pressed on.
"Analysis of variance is a special form of regression. Hierarchical linear modeling and longitudinal data analysis are based on regression. That t-test you might know and love? Well, it's really just an F-test, which is regression. When the numerator degrees of freedom equal one, F is equal to t-squared. To put this simplistically, all we are really doing in any of this is looking at variance and its possible sources."
Nothing from TheGoodOne for a few minutes. Finally he typed only, "That's kind of sad, that that's all we do."