Friday, May 8, 2009

A Semi-Defense of Psychology

My response to Debbie’s interesting comments about school and psychology took on such length that it was easier to make a new post out of it.

Either the field of psychology (or the undergraduate curriculum of the field, since there's always a lag) moved drastically in 10 years, Debbie’s professors had a weird idea about what to teach, or ... I don't know, something. Her experience of psychology courses is very different from mine.

Of course, other than intro (which is a class that perhaps doesn't necessarily have to suck, but I suspect usually does, due to being disorganized and dissatisfyingly general) and my industrial-organizational psychology course (another lower-level course), my classes did not tend to use multiple choice exams. It's easy to have a superficial understanding (as Debbie points out, from being a human among humans) that certain phenomena exist, but it's harder to cogently explain what's actually going on with them, contrast the various theories / hypotheses about them and the evidence for them, critique the methodological approaches used to support specific research findings, or whatever.

That’s one reason I dislike multiple-choice exams – they do not reward more full or nuanced understanding of the material but frequently come down to (as Debbie and her friend experienced) an ability to see through a trick, identification of a more plausible answer, or employment of other content-irrelevant multiple-choice-test-taking skills. In some cases, I suspect that these multiple-choice exams for low-level courses (perhaps in many disciplines) are used as much as a measure of whether the student read the book as anything else; even when it’s something “everybody knows” about, you probably did have to read the book to consistently recognize what specific theories / phenomena are called in the literature. For instance, you may “know” that people who have one positive trait (e.g. are good-looking) are often assumed to have other, unrelated positive traits, but if you didn’t read the book, you may not be able to identify this as being called the “halo effect.”

A particular multiple-choice question type that I hate, which is perhaps more common on exams developed from the textbook publisher test bank, is the one that tests your memory for some very specific example in the book. Something like "The story about Eldritch the Dog that opened Chapter 7 illustrated which of the following phenomena..." you might have at least a chance at, since that's a part you probably did read and are more likely to retain due to its nature as a narrative. But something like "The experiment by Rudy and Jones [completely obscure researchers who may or may not be personal friends with the textbook author or on the author's tenure committee] described in the 'Contemporary Research' box found which of the following to be true..." is tough.

I agree that the GRE subject test does appear to be oriented toward "name the theory / name the researcher" type questions, but at this point, very few programs require or even care about this exam from people with a degree in pyschology. My understanding is that it's mostly recommended for people wanting to enter a psych grad program without the undergrad degree to demonstrate some base level of knowledge of the terminology / major theories / important researchers that is needed to make sense of the literature.

For instance, I have read that one of the reasons that economists and psychologists have difficulty making sense of each other is that economists define their theories using math and psychologists define their theories with reference to other theories, using terminology that has specific meaning within the field. This justifies to a certain extent a preoccupation in psychology with this kind of “factual” knowledge about the history of the discipline, its jargon, and its well-known researchers. I mean, we have to keep those economists at bay somehow.

Maybe some people find the theories in experimental psychology more “obvious” than I do. However, I think that people in general often rely on an understanding of psychological processes that is very shallow, woefully insufficient, or downright wrong. I wouldn’t even know where to begin listing all the things people misunderstand – how memory works, what prejudice is and how it functions, how attitudes are formed and beliefs validated, the way they themselves make decisions, the extent to which perception does not capture “reality,” how to motivate others, etc. I only know a little bit about these things myself. Even top researchers in the various specialty areas do not understand these things very well.

I know that since this is my field, I am apt to be protective of it. I need to believe it to be intellectually rigorous. I need it to have something to add to human understanding beyond what your grandmother “knows” after a lifetime of observing people's behavior and watching Oprah and its ilk. No one wants to believe their own field of endeavor is nothing more than writing up obvious truths in fancy language and diagrams. But I do think that a lot of the subtlety of academic psychology is easily lost on people. Experimental psychologists can't just see some phenomenon in the world, then turn to each other and say either "Birds of a feather flock together" or "Opposites attract"; such post hoc pseudo-explanations don't get you anywhere scientifically, though most people seem happy to settle upon any neat explanation whatsoever and smugly move on.

But it’s one of the crosses social scientists have to bear: everybody thinks they already know, from their own experience in the world, 80% of the important things the discipline has to say, and they think the other 20% is wrong (or at least, may be accurate when talking about other people, but not themselves).

Thursday, May 7, 2009

In Lieu of Resume, Please Accept...

Just now I was planning to email my resume to someone and when I went to attach the file, the folder that opened was full of Leo photos. After recovering from an unexpected mega-dose of Cute, I considered how amusing it would be to revise the email to say:

In lieu of my resume, please accept the attached as an indication of my interest and aptitude for the job.

Sincerely,
Sally Porter

I am containing my enthusiasm for this job

In case this scares you - wait, I thought she was going to graduate school! I thought she already had a job! - be assured that nothing has changed. This is probably going to come to nothing, but a friend of a friend has her own research organization and she's looking for some freelance workers. She asked for my resume and I thought, Sure, why not.

Wednesday, May 6, 2009

Other People's Grades

I was really struck by this post on the LiveJournal applying to grad school forum. In a nutshell, an English/history major questions whether having taken 6-7 courses outside her majors Pass/Fail in the last 2 years will look bad on her transcript. (Oh, sorry, it's now called Pass/"No Pass" - insert curmudgeonly rant on the ridiculousness of the self-esteem movement in perverting language here if desired, or ignore.) The original poster (OP) admits to having taken this P/NP so that she could preserve a 3.92 cumulative GPA and focus extra effort on classes in her major.

I didn't comment on the post because I have nothing much to say to this particular person, and even though I have critical remarks about it, they aren't directed at this individual. This is what one's own blog is for.

First, taking 6-7 classes as P/NP in 2 years seems like a lot of P/NP. While I do not agree with the commenter who interpreted this as "slacking," it does seem to be a sort of questionable way to artificially limit one's workload. If I personally had a 3.92 GPA with 6-7 no-grade courses, I would feel very much like I have this on my transcript:

Cumulative GPA = 3.92*

*Note: 6-7 courses taken Pass/No Pass not accounted for in GPA. Recalculate cum GPA substituting C- (the lowest Pass grade) for these credits to yield real GPA of 3.53 - 3.58. Take off additional points because of how lame it is to pretend** to have a 3.92 when one doesn't really have one.
** Student may bring this transcript to Big Joe's Tattoo Shop to receive a 50% discount on "I Am An Actual Imposter" tattoo. (Students who took all math and/or econ courses P/NP may ask one of their quantitatively better prepared friends to explain the concepts of percentages and discounts.)

Second, the OP doesn't say what the courses were that she took P/NP, and I think that makes a difference in my personal interpretation of her GPA. If she took some demanding classes that people in English/History don't usually take, then the P/NP doesn't bother me much. If she took classes that other applicants to her grad programs also took, but took for a grade, then I look more disfavorably on the P/NP since it looks like she didn't think she could take those classes and do well or couldn't handle a normal workload.

Third, WTF is with a situation that could lead one commentor to note that she took a language class P/NP because her earlier "B in German...KILLED [her] GPA"? How can a single B do that even if you have been a 4.0 up to that point? If we assume she was halfway through her college career (60 hours) and had all A's plus one B, that results in a cum GPA of 3.95, hardly "KILLED." (One assumes one of the courses she got an A in was not statistics or that she has a very different idea about what it means to hurt one's GPA from mine.)

Fourth, what is with the high GPA's all these humanities majors report on this web site? I recognize that this is by no means a representative sample, but the social science and physical science majors seem to me to self-report lower GPA's than the English etc. majors. And in talking to classmates at school over the last couple years (mostly science, engineering, math, and CS majors), these people are not regularly getting a mix of A's and some B's in their classes either.

Posing this question to Robert, he found a report that Texas A&M puts out every semester with the grade distribution for every single class at the university. (He was familiar with it from when he was teaching.) And we looked up some numbers from the Spring 2008 semester. While this won't tell us precisely what GPA's look like for e.g. English majors versus Math majors, at least it will show what kinds of grades people get in the classes in these departments.

Department Total GPA

Perform. Art 3.490
Liberal Arts 3.272 (includes e.g. women's studies)
Computer Science 3.249
English 3.075
Hispanic Studies 3.072
Languages 3.070
Sociology 2.914
Philosophy 2.892
Psychology 2.885
Anthropology 2.884
Communications 2.879
Statistics 2.833
History 2.795
Poli Sci 2.783
Chemistry 2.714
Physics 2.684
Economics 2.669
Biology 2.551
Mathematics 2.320

School Total GPA

Education 3.509
Agriculture 3.167
Engineering 3.141
Business 3.118 (remember, usually must have/maintain a 3.0 to stay in the bus. school)
Liberal Arts 2.936 (includes humanities & social sciences)
Geosciences 2.903
Sciences 2.575

So these data are consistent with the claim that people tend to get higher grades in humanities courses than math/science courses. The social sciences are inbetween, which is about what I would have expected. (And yes, it also lends support to the notion that education courses are a joke. A 3.5 average?)

I also liked seeing the differences in average grades for the more advanced courses (taken primarily by majors) in math and English.

Serious-ish Math Courses (not everyone takes)

Calculus 1 2.071 n=2694
Calculus 2 2.036 n=1427
Calculus 3 2.295 n=696
Discrete 2.103 n=109
Linear Algebra 3.153 n=163
Diff Equations 2.666 n=891
Adv. Calculus 2.297 n=54
Fourier Series 2.920 n=56
Applied 2.743 n=179

Serious-ish English Courses (not everyone takes)

Am. Lit 3.025 n=459
Eng. Lit 1 2.753 n=335
Eng. Lit 2 3.033 n=190
Creative Writing 3.326 n=234
Adv. Comp. 3.176 n=135
Tech. Writing 2.861 n=556
Shakespeare 2.864 n=136
Am. Ethnic Lit 3.529 n=71
Child Lit 2.898 n=257
Adolescent Lit 3.045 n=96
Chaucer 2.564 n=57 (A tough one)
Women's Lit 3.226 n=34
Senior Seminar 3.367 n=101

So it doesn't appear to me that the relatively higher grades in English classes is due only to freshmen doing better in, say, their required Rhetoric & Composition class than their required College Algebra class. There is a tendency for higher-level classes to have higher grades in English than in Math also.

I can think of a couple of things that could contribute to The Great Humanities/Science GPA Divide:

1. It would be easier to make the course content of e.g. an English class simple (and hence easy to get a good grade in) than it would to make the course content of a math class simple. It's more obvious how to dumb down the content of an English class. After all, we all take English classes that require reading and writing essays starting in 7th grade and we don't all get F's, and I have seen literary criticism by actual professors on the same works that I read and wrote about in junior high (e.g. Great Expectations), albeit obviously at a very different level. However, there's only so easy you can make a math class and have it actually cover the relevant material. Differential equations is differential equations and 7th grade algebra is 7th grade algebra - they are not revved up or dumbed down versions of each other.

2. It is easier to resist grade inflation teaching courses like math and chemistry than it is English and history because the quantitative nature of the work makes grading more objective. (I am not saying hugely objective, simply more objective than with grading essays.) Social sciences would fall inbetween the humanities and sciences in grade inflation.

Hey, it occurred to me that the Texas A&M data might be able to provide evidence for this hypothesis. Here's what I found comparing GPA from spring 1986 (1st year in data set) versus spring 2008:

Math: 1986 = 2.286; 2008 = 2.320; change = +.034
Psych: 1986 = 2.685; 2008 = 2.885; change = +.200
English: 1986 = 2.673; 2008 = 3.075; change = +.402

Sciences: 1986 = 2.536; 2008 = 2.575; change = +.039
Lib Arts: 1986 = 2.644; 2008 = 2.936; change = +.292

This is the pattern I would have expected. Low grade inflation for math (and the sciences), high grade inflation for English, and moderate grade inflation for psychology. While the data could be interpreted as beng somewhat incomparable due to different types of students in 1986 versus 2008, I would have expected a secular downward trend in grades with time (given no grade inflation) because more marginal students are entering college in 2008 than in 1986.

Here are the average grades for the university as a whole over time (spring semester reported). Cue the Grade Inflation Alarm! Toooot toot tooooooot!

University overall
2008 = 3.051
2001 = 3.053
1996 = 2.989
1993 = 2.951
1991 = 2.874
1986 = 2.819

It seems likely to me that if, with a very little time spent looking at data on the Internet, I can find support for a grade inflation problem in the humanities (yes, from one college, but it seems like a relatively "normal" and representative one for this purpose, being a large state university), grad school adcoms in the humanities must be very well aware of what's going on with grades. No wonder I see prospective English or Women's Studies PhD applicants fretting publically about their "miserable" 3.7 GPA's and such. No wonder the original poster from the LJ forum was worried about protecting her 3.92 GPA from grades in other classes (particularly if they were from the physical sciences, where getting high C's is typical).

Ambiguity

I found the title of the blog post on Wardrobe Refashion "Resigning - For Life" sad and confusing until I realized she meant "signing up for WR again, this time for life" (the options for signing up appear to be 2, 4, or 6 months, or for life) and not, you know, resigning. I thought, Man, she must really not like making her own clothes, buying from thrift shops, and recycling clothes if she is announcing that she is out of this game forever. What would prompt such a complete repudiation of the WR ideals?

This is not quite up there with one of my favorite bits of ambiguity, though. A gazillions years ago on one of those crime TV shows, the episode opened with a view of some letters on a glass door, seen from the backwards view. It had a person's name and then the letters THERAPIST, which I parsed at first as "The Rapist." OK, the guy is a rapist - there are often rapists on crime TV shows - but he has an office for it?!

(I just checked to see whether therapist.com is a valid web site, and sadly, yes. Ah, I assumed it would be about psychotherapy, but instead it's physical therapy. At least this means they probably do not offer advice for rape victims.)

And yes, I am one of those people who see the letters NOWHERE as indicating "nowhere" and not "now here" so it's not that I am just quick to add breaks in the letters to create multiple words. Apparently that occurs when it results in the most depressing or pessimistic of my choices.

Monday, May 4, 2009

Jobs and Health Insurance - Two Great Things That Don't Work So Great Strictly Tied Together

Even in a good economy, a lot of people really fear losing their job. A loss of your job is typically accompanied by other losses such as:

* Income
* Self-esteem / self-confidence
* Self-identity
* A sense of belonging
* A venue for productive achievement
* Health insurance

Health insurance? This is basically crazy. In most cases, a person would probably not willingly bundle their income and their health insurance coverage in this way – it’s too risky. Of course, I would also expect that in most cases, a person would not choose to invest any meaningful amount of money into shares of the company they work for, but people do that a fair bit (I’m not referring to stock options here but people actually purchasing company stock outright using their own money). So maybe what I really mean is that it a bad idea and clearly not optimal for most people to have their income and insurance both dependent on their specific current job.

The development of COBRA, which allows a terminated worker to keep their health insurance coverage for 18-36 months, helps somewhat with the last of these problems because it allows you to purchase insurance at a reduced group rate rather than paying for new insurance at the individual rate. However, insurance through COBRA is still amazingly expensive because you pay the full cost of your insurance (which your employer probably has been subsizing previously) and a 2% administrative fee.

For example, the “free to me” health insurance coverage I had at my previous job would cost $365 per month under COBRA. Family coverage would cost $1067 per month instead of $344. (Note: as you can see from this fee schedule, the “state pays” subsidy for individuals is $361 per month, while the figure for families is $705 per month.) So at the same time that your income has gone down drastically with the loss of your job (in Texas, the maximum weekly unemployment benefit is $392, or about $20,000 per year, no matter what your prior income was), you undergo a new major expense.

And while it’s possible to argue that what is often called “health insurance” is really as much “access to health care, including medication and treatment for ongoing medical problems, at lower rates negotiated by the insurance company” as it is insurance against unknown future risks, people derive a lot of benefit from the negotiated rates that is lost if they don’t have insurance. But many people do worry in the absence of insurance about the economic consequences of a medical catastrophe hitting someone in their family. (Of course, the financial disaster attendant to major medical problems is bad even with insurance coverage.)

For instance, it is common knowledge that a large proportion of bankruptcies are caused by medical expenses. Even though this “fact” is disputable (how many people would have been able to pay for their medical care if they weren’t already living at the edge of or beyond their means such that any unexpected, unavoidable expense led to financial ruin? As this article points out, “all debts are fungible”), this belief contributes to the sense of fear that makes people feel hugely dependent on their current job to protect them from economic ruin.

It’s not good for people to feel so dependent on their job for a lot of obvious reasons, including sticking with a job they should leave, being too afraid to ask for changes to their job (a raise, educational opportunities, etc.) or rock the boat in any way, becoming depressed and/or resentful about being trapped, and overgeneralizing unhappiness in a single job such that they believe they hate all possible jobs or all possible lines of work.

But it also can easily lead to increased populist demands for job “stability” or “protection” that may sound really beneficial and practical but ultimately could undermine the flexibility of the market. People are not very thoughtful about this even in good times, but when you’re afraid of losing your job and unemployment has gone up, it’s hard to see that change is a good and necessary part of a functioning market. Creative destruction just looks threatening. So much more appealing is something like the Dave idea that the president can just guarantee a job for every American who wants one (perhaps especially when espoused by the charming Kevin Kline).

I think the American labor market is already too inflexible. While it is naïve to assume that everything you learned in your undergraduate microeconomics course about how people elect to work the number of hours where the continuous supply and demand curves meet, it does strike me as unnecessarily screwed-up the way that there is this categorical divide between full-time (frequently defined as 40 hours+ per week) and part-time work.

Particularly when you look at the market for professionals / white collar / office workers (I don't have enough familiarity with what things are like for people in manufacturing, retail, etc.), people have a limited choice in the number of hours they work. Robert makes a good point that to a certain extent, people can choose their work hours within the range of 40 and up. Different careers, industries, and employers have different cultures/requirements about work hours. (Of course, this provides only very broad flexibility and assumes that your interests in work hours and type of work line up well.)

However, if you are interested in a workweek that falls below the 40 hour mark, you probably will need to content yourself with lower-level, lower-pay jobs (generally without insurance benefits). And even if you were willing to take the economic hit (which a lot of people aren’t), if you are a knowledge worker, it can be difficult to find part-time work that makes use of your abilities. So there seems to be a bit of a double-whammy in that a skilled person may easily find herself in a part-time job that both pays relatively poorly compared to the comparable full-time job and that is below her skill level. For instance, a market researcher might not be able to find a job doing survey design and analysis but instead may get hired as a part-time interviewer. Nobody benefits from this.

Obviously, there are other forms of fixed costs / overhead associated with having two people working 30 hour weeks versus one person working 60 hour weeks and efficiencies that can potentially be gained from having a smaller number of workers to coordinate (though at what point does an individual’s inability to continue working at high productivity with the physical exhaustion of working more hours and mental exhaustion of keeping so many different projects afloat kick in?). But it seems to me that the current system, with a large amount of compensation going into benefits such as health insurance, provides further incentives for employers to get as much work out of as few people as possible and shuts people who do not want or cannot commit to working 40+ hours per week out of jobs that meet their skill level and they could excel in.

Is this all to say that I support nationalized health care? Not exactly. But I do think that there are serious potential gains from decoupling people’s jobs and health insurance coverage.

Friday, May 1, 2009

New Packaging

Kraft shredded cheese now comes in a new package and a new size, 7 oz instead of 8 oz. While I'm sure 99% of people who notice this are going to be annoyed by the downsizing (evil marketers! reducing the quantity and trying to fool me! why don't they just increase the price...wait, I don't want them to do that either! raising prices in unAmerican! why isn't Barack Obama stopping this price gouging! doesn't he know there's a war on! wait, I don't want to sacrifice anything! doesn't he know there's a recession on!), I'm not.

My first thought was, All those recipes that call for 8 oz of shredded cheese will now get 7 oz; if I save 1 oz of (2% milkfat) cheese over a dish that makes 4 - 6 servings, that's saving me 13-20 calories per serving, and I probably won't be able to tell the difference at all.

20 calories may not seem like a lot, but "it is estimated that 80% of the population gains weight because of a calorie excess of less than 50 calories a day" (Wansink and Huckabee, 2005).

Since this downsizing is occurring against a backdrop of increasing serving sizes in recipes, restaurants, and "giant" candy bars, boxes of cereal, trail mix, potato chips, etc., in stores - which are culprits in the obesifying of America - well, we need all the help we can get. I choose to view Kraft's decision to decrease the amount of cheese in the package as an inadvertent aid to me in my goal of reducing my calorie intake as painlessly as possible.

I mean, slightly decreasing my cheese consumption beats all hell out of eating compressed timothy hay pellets for breakfast, as dictated by the Rabbit Diet.

Source:
Wansink, Brian and Mike Huckabee, 2005, De-Marketing Obesity, California Management Review, 47(4): 7-18.

Dance For a Lifetime

This morning, I enjoyed reading this article in The Economist about 90-year-old Merce Cunningham, renowned American dancer and choreographer, whose new piece "Nearly Ninety" opens today in Madrid.

More background on Cunningham here. I thought this was interesting: his first solo concert was performed with musician John Cage in 1944, and like Cage, he incorporated the element of chance (flipping coins, etc.) in creating his pieces.

Even though he's old and in a wheelchair, he's continuing to work (although the article suggests with a greater level of input from his collaborators than previously). To me, that's pretty inspirational.

It was also an appropriate thing to see today as my sister celebrates her last day at her tech job and takes on dance as a full-time pursuit.

Break a leg, Jen!