Do Professors "Cause" Anything?
We already know that Professors write esoteric papers that few read. Today's New York Times now tells us that we don't influence our students' ideology (see Do Professors Cause Obamania? ). I guess that we have no choice but to sit at the beach and think positive thoughts.
There is an obvious weakness in today's NYT piece. Students self select what University to attend and once they are at a University, they choose what classes and what professors to take. We need some clever political scientist to think of an instrumental variables strategy to randomize the ideology treatment.
Jim Heckman would counter that even if this "clever" political scientist came up with such an randomization device (perhaps differential tuition scholarships for people from under-represented states?), the students may still drop out of the class or not attend if they are shocked by what they hear. Intuitively, they may chose to not take the treatment. This is one of Heckman's points about essential heterogeneity. If I am a big liberal (and I know that I'm a big liberal while the researcher knows nothing about my core ideology), I may not be willing to sit through 13 weeks of Harvard's Harvey Mansfield's Manly lectures on the joys of Harvard in the 1950s and 1850s.
www.people.fas.harvard.edu/~hmansf/
So the researcher would observe liberal students selecting liberal classes (the 1960s) taught by liberal professors and conservative students selecting classes at the other end of the spectrum. The researcher would never observe the "counter-factual" of how does student ideology evolve when they are "treated" with an alternative perspective.
There is also an issue of re-enforcement. If you take a conservative class but all of your friends are liberal, what happens in the dining hall as you debate what you have learned? There may be interaction effects between ideas from Profs and from peers.
In a diverse society, who is the "marginal" student who actually takes a course that challenges his/her world view? The Angrist/Imbens LATE approach would say that you could answer the political scientists' causal effects of ideology question for the subset of students at the margin who due to a randomization took a course they usually wouldn't take. Have the political scientists been this subtle? I hope so.
There is an obvious weakness in today's NYT piece. Students self select what University to attend and once they are at a University, they choose what classes and what professors to take. We need some clever political scientist to think of an instrumental variables strategy to randomize the ideology treatment.
Jim Heckman would counter that even if this "clever" political scientist came up with such an randomization device (perhaps differential tuition scholarships for people from under-represented states?), the students may still drop out of the class or not attend if they are shocked by what they hear. Intuitively, they may chose to not take the treatment. This is one of Heckman's points about essential heterogeneity. If I am a big liberal (and I know that I'm a big liberal while the researcher knows nothing about my core ideology), I may not be willing to sit through 13 weeks of Harvard's Harvey Mansfield's Manly lectures on the joys of Harvard in the 1950s and 1850s.
www.people.fas.harvard.edu/~hmansf/
So the researcher would observe liberal students selecting liberal classes (the 1960s) taught by liberal professors and conservative students selecting classes at the other end of the spectrum. The researcher would never observe the "counter-factual" of how does student ideology evolve when they are "treated" with an alternative perspective.
There is also an issue of re-enforcement. If you take a conservative class but all of your friends are liberal, what happens in the dining hall as you debate what you have learned? There may be interaction effects between ideas from Profs and from peers.
In a diverse society, who is the "marginal" student who actually takes a course that challenges his/her world view? The Angrist/Imbens LATE approach would say that you could answer the political scientists' causal effects of ideology question for the subset of students at the margin who due to a randomization took a course they usually wouldn't take. Have the political scientists been this subtle? I hope so.
posted by Matthew E. Kahn at 7:32 AM
0 Comments:
Post a Comment
<< Home