Eternity lasts a very long time. Our resources, though “infinite,” are not unlimited....Focus groups have suggested that offering a mere year or two of heavenly bliss, coupled with the threat of a single hour spent bathing in hot pitch and being harassed by demons, would generate ninety-seven per cent of the current program’s salutary effect on mortal behavior. (Interestingly, eternity itself is now perceived as a disincentive by blessed souls with more than two years of college education.) This suggests that severely scaling back Our incentive plan—and its attendant costs—would not lead to a significant diminution in faithfulness, obedience, repentance, or other benefits accruing to Ourself.
While you were watching the NFL playoffs this weekend, economists converged on New Orleans for the American Economic Association's annual meeting - think Mardi Gras, but without the fun.
For the early risers yesterday morning, there was a panel called "Student Incentives in Action: Experimental Evidence from Offering Money for Educational Achievement." Roland Fryer and Ceci Rouse, originally scheduled to present, were no shows, but there were three other papers presented: the first by Case Western's Eric Bettinger called, "Paying to Learn: The Effects of Financial Incentives on Elementary Test Scores" (policy brief available here), a second by Josh Angrist, Daniel Lang, and Philip Oreopoulos, "Lead Them To Water and Pay Them to Drink: An Experiment With Services and Incentives for College Achievement," and a third by Angrist and Victor Lavy called "The Effect of High-Stakes High School Achievement Rewards: Evidence from a Group Randomized Trial" (the latter two are available here).
Bettinger's study was based on a randomized experiment where students were paid for performance on periodic math, reading, writing, social studies, and science tests. These incentives increased test scores only in math, but not in any other subject. And the kids who gained the most from receiving the incentive were those already performing at higher levels, not the lowest performing students. Here's the kicker: The study was multi-year, such that some students were given incentives in one year and not in the next. Advocates of incentives argue that while students will react to the cash at first, when the incentive is taken away, they will learn "for learning's sake." Yet Bettinger found no carry over effects when the incentive was taken away, writing, "This may suggest that the existence of external motivation has a negative effect on the intrinsic desire to learn." What's worse, kids reverted back to their initial achievement level, suggesting that the incentives affected not permanent learning, but short-term effort.
Bettinger also shared two funny stories about how teachers used the cash as motivation: in one case, a teacher had the kids chant "Show me the money!" In another case, a teacher hung a giant $100 bill in her classroom. For writing practice, kids were asked to write about how they would spend their money.
In the other K-12 paper, Angrist and Lavy found that offering financial incentives to pass a high-stakes test in Israel improved outcomes for girls, but not boys. The effects on girls were largely driven by an increase in passing rates among those who had a relatively high chance of passing these exams to begin with.
Taken together, these studies don't bode well for the current drive to improve outcomes for the lowest performing students by paying them.
