Friday, October 16, 2009

Classrooms and Political Engagement

An "open classroom environment" can have positive effects on adolescent "civic knowledge and appreciation of political conflict," according to this study in the academic journal Political Behavior.

Okay, the first question from you budding methodologists is -- what's an open classroom environment? 

Table 1 of the study (page 443) outlines what is basically a classroom environment where kids feel comfortable asking questions, expressing opinions, and where teachers respect student opinion (or at least, I suppose, don't beat the crap out of the brats for mouthing off).  But yeah, I can buy this.  So does it make a difference?

Table 2 includes a monumentally long regression analysis to support the argument.  Classroom environment, even after all these statistical controls, does explain some unique variance -- PhDweebspeak for it is still a significant factor despite controlling for lots of other explanations (beta a mere .06, but significant despite all the other controls).  So what predicted civic knowledge among kids?  Classroom environment, obviously, but also significant factors were: expected education (how far you think you're going to go on in school), race, reading of books, discussing politics at home, and perception of the classroom environment, which is apparently different from actual environment.

What drops out as predictors of civic knowledge?
  • Sex.  That's interesting given sex becomes a predictor for adults, with men scoring higher than women on tests of political knowledge (I've blogged on this research here and here).  But at this early age, no gender effect.  So the differences may emerge later.  Worthy of further investigation.
  • News Media use.  At this age, less surprising.  Kids, even older ones, don't consume all that much news.  More meaningful would be the media habits of parents or guardians, but the data probed student use of media.
  • Income and Free Lunch.  These essentially measure the same thing, which makes me worry about multicollinearity.  You can't get a free lunch at public schools unless your family is under the federal poverty level, so these two variables measure basically the same thing.  I would have taken out one from the model, gone with just the other. Or perhaps combine them in some way.
  • Social studiesGasp!  Taking social studies has no effect?  Bad news for those who hope classes like these, which have largely disappeared due to No Child Left Untested, will save the day. 
In all, an interesting peek at kids versus the jillion of studies that examine U.S. adults.  I think there is some stuff above that make for interesting follow-up studies, in particular a further investigation of the sex differences in political knowledge scales.  Some good alternative explanations for that one have emerged of late (the kinds of questions asked, social learning, etc.).  More needs to be done.

No comments: