Alex Cohen is the Director of Learning and Evaluation for the Richard M. Fairbanks Foundation.

It’s well known that good jobs often require a strong understanding of science, technology, engineering and math (STEM) fields. Yet K-12 performance in STEM lags in Indiana, driven in large part by substantial disparities across race and income.

In 2017, only 51 percent of Indiana students passed the math portion of the ISTEP+ in grades 3-8. The pass rate for students with free- and reduced-price lunches was just 36 percent. And the pass rate for black students was 25 percent—less than half the rate for white students.

Raising STEM achievement, particularly for low-income and minority groups, requires investing in approaches with the highest likelihood of success.

Three recent studies point to promising approaches and add to our evidence base on what works and what doesn’t when it comes to STEM education.

There are other approaches that have strong evidence, too.

The What Works Clearinghouse, for example, highlights several effective interventions for math and science, including specific school models and alternative teacher preparation pipelines. Early childhood interventions like high-quality Pre-K and Nurse-Family Partnership have also been shown to improve math achievement, among other outcomes.

As more and more focus is put on STEM, it’s natural for enthusiasm about certain approaches to outpace evidence for their effectiveness. But focusing on programs and interventions that have been subjected to rigorous evaluation—and applying guidelines for evidence like those recommended by the U.S. Department of Education—raises the likelihood that our efforts will lead to demonstrably positive improvements in outcomes for students and help close disparities in STEM achievement.