Ever since the Texas House passed a controversial bill last month that will require school districts in the Lone Star State to spend 1 percent of their budgets on incentive-based pay, Texans have been fiercely debating the merits of merit pay.

Everyone seems to agree on one thing: the largest experiment with merit pay in Texas is not working. Last year, 80 percent of Houston's teachers qualified for merit bonuses, courtesy of that district's "school-wide" policy of rewarding teachers in schools for meeting standardized test goals. Most teachers got the same sum, a whopping $440.

A plan proposed by Texas business leaders last week sounds somewhat more promising. That proposal creates graduated rewards for the top 30 percent of teachers who boost test scores in high poverty schools. Teachers in the top 10 percent would get $7,500, and teachers in the top 20 percent and 30 percent would get $5,000 and $2,500 respectively. This plan lines up better with emerging principles of good merit pay: first, providing awards that are substantial enough to influence behaviors and second, distinguishing exceptional performance by individual teachers, not just school-wide performance. However, the plan's full reliance on test scores as the only criteria of teacher effectiveness will make it more than a little tough to implement for high school teachers. Combining evaluations with student achievement gains may be a more promising and fair approach.