Wednesday, November 19, 2008

Community College, Will You Marry Me?

A colleague sent me this article with the following comment:

I hope the data described below trouble you. they damned sure trouble me. yeh, I know, we’re trying and we’re doing better. that’s nice, but pathetic.

The issue is “student engagement” at community colleges. The question, I think that needs to be addressed but never is, is what is a realistic expectation for student engagement? Certainly we want all to be optimally engaged. Realistically, I think that's ridiculous. The data discussed in the article seems to fall into two categories: Engagement in the classroom and engagement out of the classroom. What the author says about engagement in the classroom confuses me:

“In the classroom, engagement figures were…still notably low. Twenty-eight percent said they had either “often” or “very often” made a class presentation, and 46 percent said they had either “often” or “very often” worked with other students on projects during class… These figures were higher for students at four-year institutions, according to this year’s NSSE report — 33 percent of first-year students had “often” or “very often” made a class presentation and 60 percent of seniors had done the same. Additionally, 43 percent of first-year students at four-year institutions had worked with other students on projects during class either “often” or “very often,” while 47 percent of their senior counterparts had done the same.”

I’m not especially confident that “making classroom presentations” and “working with other students on projects during class” are worthy measures of “student engagement” in the first place, but I can play along and pretend. Then the question becomes what comparisons should we make? If we’re going to compare CC students to 4-yr students, and if we’re given the option of comparison between first year students and seniors at 4-yr schools, (unless someone can present an argument otherwise) it sure seems wise that we compare CC students to first year students. So, yes, 33% making class presentations is a bit higher than only 28% - but not drastically higher (BRCC is at 22.2%). But, no, 43% is not higher than 46% - CC students are *more likely* than first year students at 4-yr schools to do group work in class (BRCC is at 44.8%). Given the data, it just does not seem to me that CC students are significantly (in a non-stats sense) less engaged in class that their 4-year counterparts.

Outside of the classroom, yes, there seems to be a big difference. 16% of CC students “reported that they discussed ideas with their professors outside of class.” (BRCC is at 10.7%). While it isn’t in the IHE article, I dug up that “nearly 2/3 of 1st year students…at least sometimes discussed ideas from readings or classes with faculty members outside of class” (page 11). Sure, that is a significant difference. But I’m simply not convinced it is realistic to expect similar “out of class” engagement numbers between CC and 4-year students. 62% of the subjects in the CC data are part-time. It never quite mentions the percentage in the 4-year data who are part-time, but it does indicate that at least some of the analyses are based only on full-time students (page 18). I think it unrealistic to expect that part-time students spend as much time out of class “engaged” with academics as full-time students.

I suspect also that a lot of the variability between CC's on "out of class" student engagement is a function of whether the school is commuter (like BRCC) or residential (with dorms, like those I worked at in Colorado & Arizona). It always seemed to me that students spend a lot more time just “hanging around campus” when they are (1) full-time, and (2) at a residential college (CC or 4-yr). More time “hanging around campus” likely corresponds with more time chatting with professors out of class. Realistic expectations should be different for this measure of "engagement" at different types of colleges.

Oddly, the IHE article doesn’t even mention any data on the two measures of “student engagement” that I think are really indicative of it: Student effort and academic challenge (anecdotal comments are made on academic challenge).

Ultimately, to get at my colleague's point: No the data don't trouble me much, nor do I think them sufficient to conclude that what we're doing is pathetic.

Thursday, November 6, 2008

You sexy correlational bitch!*

So it looks like teenagers who watch a lot of sexual themed TV are more likely to get pregnant. Let's see what the media says. First, here's the actual researchers:

The study found that frequent exposure to TV sexual content was associated with a significantly greater likelihood of teen pregnancy in the following three years.

This, obviously, is a simple correlation (with some other variables controlled, thus likely some variation of a partial or semi-partial, but still, simple correlation): As Variable 1 (amount of TV sexual content viewed) increases, Variable 2 (likelihood of pregnancy) also increases. Will the media describe the findings as correlational? Here's the NYT:

Shows that highlight only the positive aspects of sexual behavior without the risks can lead teens to have unprotected sex (emphasis mine)...

Ah yes, one of the most basic science errors (lies?): Variable 1 is correlated with Variable 2, thus Variable 2 leads to Variable 2. Damnit, man! Anyone who has ever taken a freshman level social science class knows correlation does not mean causation (although they may not know, as I didn't until just now, that this was called the cum hoc ergo propter hoc logical fallacy)! How did NPR do?

Still, the authors say the study has limitations — that they can't rule out other factors that may influence the findings. For instance, it's possible that teens with advanced sexual attitudes are more likely to seek out more TV shows with sexual content.

Good job, NPR. There are plenty of other reasons why Variable 1 might be correlated with Variable 2 aside from Variable 1 causing Variable 2. This shouldn't be hard!
*The title is for you, Leslie.

Monday, November 3, 2008

“The outrage that the F.C.C. pretends to feel is false”

I often use emotional responses to swearing as an example of Albert Ellis’s approach to mental health (he was, after all, “the first psychologist ever to say ‘fuck’ and ‘shit’ at the American Psychological Association conference”). Anyway, Ellis argues that our emotional responses are not about particular events (like swearing) but are a function of the way we interpret events (for example, I will respond differently to swear words if I decide to think differently about them). Yet again, the FCC is trying to decide if saying “fuck” and “shit” on live TV is always punishable, and thus, censorable (or vice versa). One judge says that even if Bono simply says that winning an award is “really, really fucking brilliant” it is still about sex!

(W)hatever the speaker’s intentions… “a substantial part of the community, and of the television audience, will understand the word as freighted with an offensive sexual connotation.”

But if that is the criteria for censorship, how long before people can no longer say “69” on television? I would guess that “a substantial part of the community, and of the television audience, will understand the word as freighted with an offensive sexual connotation.” But if I simply count, “67, 68, 69, 70” and you think simultaneous oral sex, I’m sorry, but you’re the pervert.