This is a bit old, but in case you haven’t seen it…
A few weeks ago, Jake wrote a post about the importance of teaching during grad school. I couldn’t agree more–some of my best experiences in grad school to date have been in front of a classroom of keen undergrads, their young minds yearning to be filled with the minutiae of microeconomic theory (or so I have convinced myself). I find it absolutely true that having to teach a concept cements it in my mind, and it is my fervent hope that a few of my students will at least consider becoming economists.
There is also, of course, the question of what the students get out of classes. A while back, one of my classmates sent around an AER paper by George Borjas entitled “Foreign-Born Teaching Assistants and the Academic Performance of Undergraduates“. The gist is that undergrads who have foreign TAs perform worse on their exams than their classmates with “native-born” TA’s.
This raised a few eyebrows among grad students at the LSE, where just about everyone is foreign. I can think of one British grad student in the econ department. Presumably, the issue in question is one of fluency in English rather than actual “native-ness”, so I should also count one Australian and a smattering of Irish and North Americans. That still leaves the vast majority of economics undergrads being taught by TA’s for whom English is a second (or third, or fourth, or fifth–some of my classmates are amazing) language.
So just how bad is this for them, and why?
The difference Borjas found is small, but it is statistically significant. Students who have foreign-born TA’s score roughly 0.2 points lower on the usual 4-point grading scale (A = 4, A- = 3.7, B+ = 3.3,…). And this result is robust to several specifications, including controlling for overall GPA, which serves as a proxy for general academic ability.
Borjas looked at an introductory-level course sequence in economics (micro and macro) at a large public university in the US. All students who pursue an economics major must take this sequence, which involves lectures and weekly sessions with TA’s. They have different TA’s for the micro and macro component. Borjas surveyed undergraduates enrolled in the micro component in three separate years, asking them whether their TA’s in the component courses were foreign, their final grades, their GPAs, and other personal characteristics. He also asked them to comment on whether foreign-born TA’s had better or worse communication skills than native-born TA’s, and on whether foreign-born TA’s were better or worse prepared for class. The students generally reported no difference in the level of preparation for class, but almost 80% of students said that foreign-born TA’s have worse communication skills than native-born TA’s.
Table 2 shows the results of Borjas’s regressions:
This is a fixed-effect regression (columns i, ii, v, vi, vii), which means that a dummy variable (1 if something is true, 0 otherwise) for each student was included. The individual student fixed effects control for any student-specific differences (such as ability or effort) that might affect their final grade but don’t vary between the two components (micro and macro). So the coefficient you see on “Foreign-born TA” represents the isolated effect of having a foreign TA. Looking across specifications, the -0.2 coefficient doesn’t change much depending on what other control variables are included, which indicates a pretty solid effect.
Look now at column (vii).
Column (vii) interacts the foreign-born TA dummy with the undergraduate’s perception of the class-preparedness of foreign-born TA’s. The evidence indicates that foreign-born TA’s do not worsen the scholastic achievement of undergraduates if they are better prepared than native-born TA’s. It seems as if additional class preparation may resolve the teaching difficulties encountered by foreign-born TA’s.
So spending a bit more time preparing the week’s lesson ahead of time ameliorates the communication problem.
Finally, Borjas checked to see if there was a differential effect on American versus foreign-born undergraduates. Looking at columns (i) and (ii), in which Borjas ran the primary regression separately for American and foreign students, it turns out that foreign-born TA’s hurt the performance of their American students, but have no adverse effect on their foreign students. Borjas reckons that this is because most of the foreign undergrads were Asian, and so were most of the foreign-born grad students.
The evidence suggests that foreign-born graduate students do not have an adverse impact on the academic achievement of undergraduate students who are “like them” (perhaps both in terms of language and culture) but do have an adverse impact on undergraduates who are sufficiently different.
I have an additional hypothesis: even if foreign-born undergrads are being taught by TAs who are “like them”, they are still being taught in English. They therefore may come into the course expecting to have to work a bit harder to understand the material as it’s being presented and this added effort may compensate for a possible linguistic disadvantage. I wonder if exerting a little extra effort might not help American students lessen the adverse impact on their learning.
It’s important to bring up the question of external validity of this study. Borjas looked at 309 economics students in a single university in the United States, so it’s certainly not clear that this is a universal effect. However, given that more and more people are coming from all over the world to do their graduate work in the US (which we absolutely want!), it’s worth looking at any unexpected results of this trend. My reading of the paper suggests that, if there is an effect, a little extra effort on both sides to improve understanding and communication can solve the problem.