The New York Times reports studies that endorse streaming school classes by abilities. .

As summarised by NCPA, the article says:

"Empirical data has tended to show that students gain when they are grouped according to their skill level:

• Last year, [Dartmouth College economics professor Bruce] Sacerdote and two fellow economists analyzed students at the U.S. Air Force Academy. They found that the students largely benefited from their peers, but those benefits disappeared when cadets of the highest and lowest abilities were grouped together.
• In a 2009 paper, Sacerdote tracked Hurricane Katrina refugees across different schools, concluding that students with high abilities benefited the most from high-ability peers.
• Examining primary schools in Kenya, another study found that all students — not just the best learners — benefited when they were grouped into different classrooms according to their abilities.
• Analyzing data from one North Carolina county, researchers Caroline Hoxby and Gretchen Weingarth determined again that students benefited when they were surrounded with students of similar abilities.

I hhaven't tracked through all the links in the short NYT piece, but I wonder if the mechanism at work is similar to the mechanism that results in girls in girl's schools thriving away from the testosterone fuelled dominance of boys. Perhaps being made to feel too humble may permanently limit ambition and the confidence to try.

The NCPA summary goes on:

"The benefits of ability grouping, not just for the talented students but for all students, indicate that [current left and union endorsed plans to eliminate] gifted and talented programs could have serious, negative consequences for bright students from low income families". .