This article has too much story fluff. tldr: ------------------------------ The experiment: pay kids for performance and measure improvement in skills with a

This article has too much story fluff. tldr: ------------------------------ The experiment: pay kids for performance and measure improvement in skills with a standardized test. Four payment schemes were chosen, to be implemented in four different cities. Schools within the city were randomly assigned to the control or experimental group. In New York, 4th graders were paid $25 for good test scores, 7th graders $50. The article does not say how frequent the tests were, but the average student earned ~$180 over what looks like a school year. No effect on scores vs. the control group. In Chicago, 9th graders were paid "$50 for each A, $35 for a B and $20 for a C, up to $2,000 a year," and half of their income went into an account redeemable at high school graduation. The kids got better grades compared to the control group, but their skills--as measured by standardized test--did not improve. The article is vague about Washington. It says, "In Washington, middle schoolers would be paid for a portfolio of five different metrics, including attendance and good behavior. If they hit perfect marks in every category, they could make $100 every two weeks." The kids "did better on standardized reading tests." In Dallas, second graders were paid $2 for every book they successfully read and answered a quiz about. The kids did significantly better in the standardized reading-comprehension test, and a year later--after the rewards stopped--their scores were still higher compared to the control group. ------------------------------ That's an awful lot of variables they changed from city to city. Off the top of my head, I see - Age of the students - Whether we're rewarding performance, grades, or improved behavior - Frequency of the reward - Delay between achievement and payment - How much direct control the student had over their performance - Level of initiative in student participation It strikes me as unsurprising that the Dallas experiment was successful. The reward was for drilling the key skill (reading comprehension) opposed to an auxiliary like grades or behavior. The payment was immediate (kids are bad at delayed gratification), opposed to days or even years later. Students had good control over payoff for their efforts; reading a book and understanding it is concrete, while studying for a test when you don't know exactly what's on it is abstract. And the whole thing driven by initiative on the part of the student, instead of an administratively mandated test schedule. Those choices all seem to me to be better than the alternatives. I do think the immediate feedback for solving concrete problems is important. Kids are like puppies. They need feedback FAST for something they UNDERSTAND and can CONTROL, or the lesson is greatly diminished. If I were to generalize the success, I would recommend a standing offer to pay students to drill skills directly by solving concrete problems, with payment to be made immediately upon problem resolution. You could scale it down to a penny per trivial arithmetic sum, or up to $200 for solving any of a list of independent research or engineering challenges. I also find the success in Washington particularly interesting. I'd speculate about kids learning to respect school or secondary effects from an improved learning environment, but . . . really I wish I could see their metrics.