With Democratic presidential hopefuls floating plans for large across-the-board salary increases, teacher pay is thankfully back in the news. While the candidates focus on the critical need to raise low base pay, our drum is beating for both better base pay and much more strategic use of compensation dollars by school districts to purchase what they most need and value.

One of the obstacles preventing these dual needs from getting a good airing is that there are a number of research studies that purport to examine the benefits of strategic pay but that didn't find much of a payoff (such as here, here, and here). We would assert that while the research was largely sound, the interventions themselves suffered from all sorts of fundamental design problems, such as expecting woefully small pay differentials to motivate any teacher to do something they might not otherwise do or to draw new talent into the teaching pool, or insufficient information to teachers about how to earn incentives.

That's why these two new studies finding real benefits from targeted pay—largely, we suspect, because both efforts were so well designed—come as welcome news.

Georgia is one of 15 states now offering a differential pay system targeted to fill gaps in the supply of teachers in shortage subject areas, which often include math and science. In a recent CALDER working paper, economists Carycruz Bueno and Tim Sass put this program under the microscope and find that the supplemental pay provided to STEM teachers by virtue of this program reduced teachers' exit from the classroom

Under the Georgia program, districts start new STEM teachers at Step 6 on the statewide salary schedule, but then keep them there for the first six years of employment. That means the differential between what they would have otherwise received is highest in the first few years (about $4,500 per year) and no different by year six. Not surprisingly, attrition rates fell by as much as 28 percent in the first few years. Even in later years, although attrition increased over time, it remained lower than it would have been without the higher salary.

Bueno and Sass note that the financial incentive had no impact on converting undergraduate secondary math and science education majors into public school teachers—which hardly seems worth pointing out since these individuals had, in effect, already made a career choice to teach. The more important question of the program's ability to attract non-education majors is left off the table. (The authors did tell us they intend to answer this question in a revised study they are working on now.)

Heading further south to Louisiana, economist Ozkan Eren of University of California, Riverside analyzed Louisiana's TAP model, a comprehensive performance-based pay program in high needs schools that doesn't get nearly enough attention. This study finds a strong association between the state's implementation of TAP and gains in students' math achievement. Here too, many other performance-based pay models have suffered from basic design issues such as 'free rider' effects (school-wide bonuses might allow all teachers in a school to benefit off the hard work of a few of their peers) and a lack of formative feedback. TAP's hybrid approach of providing both individual and school-wide bonuses, all premised on an appropriate mix of student learning gains and observation feedback, offers precisely the individual incentives and feedback that are designed to counteract these issues.

Wondering if your state supports strategic compensation? Check out our guide to state policy and promising practices. Also, our District Trendline provides a scan of large districts' differential pay models across the nation.