One large feature of the education debate is the perception that richer students can "buy" increments in their college entry examinations through extra spending by their parents on private tuition. One issue with this argument is that there may be a selection effect in to private tuition with more motivated and exam-focused students selecting themselves into taking tuition. Also, the fact that their parents can afford and are willing to pay for tuition will generally correlate with other potentially beneficial background factors such as peer expectations. In the extreme case, the tuition does not add anything at all. The best students simply select this and this becomes self-reinforcing over time. On the other hand, the very best students may not need it or want it so this could lead a simple analysis to potentially underestimate the effect.
As far as I am aware this has not been tested in Ireland but a recent IZA discussion paper (below) tests it for Turkey. Their results are not glowing in terms of the effectiveness of private tuition on placement, actually finding a negative effect in some of the tighter models. Their overall conclusion for Turkey is that large amounts of spending can have an effect on placement in to college but that the decision to take tutoring itself does not have a positive effect. How this applies to other countries is clearly difficult to tell but the idea of taking out the selection bias before making an overall assessment of the effectiveness of private tuition is a good one.
Does Private Tutoring Payoff?
by Ayfer Gurun, Daniel L. Millimet
(August 2008)
Abstract:
We assess the causal effect of private tutoring on the probability of university placement in Turkey. We find that tutoring increases the probability of being placed in a university when non-random selection is ignored. Moreover, among those utilizing private tutoring, greater expenditure on tutoring is also positively associated with university placement. However, we find evidence of positive selection into tutoring, but negative selection into greater expenditures among those receiving tutoring. Accounting for this pattern of non-random selection, we conclude that private tutoring has a negative causal effect on university placement overall, but conditional on receiving any tutoring, spending more on tutoring has a positive causal effect on university placement.
Text: See Discussion Paper No. 3637
This raises the obvious question of the effect of the "grind schools". How much of their "effect" is self selection? There might also be a placebo (?) effect: if your parents are forking out cash for Leeson Street you might work your butt off to convince them that it was worthwhile or alternatively you take it easier thinking that you are bound to well given the extra tuition.
ReplyDelete