Tuesday, March 9, 2010

Wheels 'n Tests

It was so much fun going to the shop a couple weeks ago that my car went back again this week. It had died again in my parking lot, and after being jump-started, it died again after about 3 minutes. Luckily, I was driving around the complex and hadn't gone out to the street because it was sort of hard to control the car after the "power" aspect of power steering and power brakes was no longer operating. $700 and a few days later, my car has a new alternator (and the windshield wiper assembly installed and working). I drove around for about 15 minutes after picking up the car (even though the dealership is 2.7 miles away) just to see if it would keep going. I was actually pretty scared when I left the dealership and it didn't go away at all while I was driving. I guess it's going to take some time before I trust the car to both start and keep going although on a rational basis I do think that the car's problems are fixed.

Last week, I took my two remaining midterm exams, which were strikingly different in format.

The first one was an essay exam for which we had been given about 5 sample essay questions and told that if we were prepared to write a concise but meaty paragraph for every empirical article we read and knew how to answer the sample essay questions (which covered both topics and issues in the field), we would be ready to take the exam. It was a ton of work preparing for the exam, but I was able to (barely) finish in the 2 hours, 15 minutes we were given with good to excellent answers for each section. There was no time for much thinking during the exam, so the pre-exam prep was critical. I also believe that the prep for this exam, which required integration of a lot of material and organizing a lot of disparate evidence, really made me understand the information in much greater depth.

The second one was a dreaded multiple choice and short answer exam. The week before the test, he told us it was multiple choice but someone pointed out that his syllabus stated short answer, so he agreed to split it between the two formats. As I've discussed before, I think MC exams can be harder than essay exams because you have to answer specific questions with potentially ambiguous or confusing or whatever wording (and you may be required to know specific facts) rather than build your own case that you know important things about the topic. I struggled with about 4 of the MC questions, but was in good company - after the exam, several of us talked about these same questions and how we weren't sure what the questions and/or answer choices meant.

One interesting thing was that our cohort's overachiever and I seemed to ultimately interpret these confusing questions/answers to mean the same thing and answered them the same way. So it's possible that the ambiguous-seeming MC questions were low in face validity but were effective in the psychometric sense in that people who know the material well end up answering them in the same way and people who know the material less well answer in a different way. But students are very accustomed to having exams for a class have high content and face validity - to look like they are covering the material of the course well and to appear to be measuring knowledge of the material in a straightforward fashion.

It would be amusing if the exam accidentally turned out to be extremely valid but in the empirical rather than rational sense because empirical and rational methods of test construction were part of the material that we read about (but were not tested on). Rational test construction is what professors typically do - ask questions that are "directly, obviously, and rationally related to what the test developer wishes to measure" (Funder, 2007). Empirical test construction involves finding questions whose answers differentiate groups of people. A classic example of this kind of test is the MMPI, used to separate people with a higher likelihood of psychopathology from the rest. For example, the question "I sometimes tease animals" tends to be answered false by people with depression.

But bottom line: I'm happy to have my car back in play and those exams behind me. Now I need to seriously buckle down on my research.

No comments: