This commentary was originally published by The Hechinger Report.

Research has shown both short- and long-term benefits of high-quality pre-K. So it was surprising when a recent study found that children who attended Tennessee’s state-run voluntary pre-K program actually performed slightly worse on sixth-grade tests and behavioral measures than children who were wait-listed for the program.

This single study has renewed the debate about the value of universal pre-K and the need for continued investments in early learning.

Yet, as noted by researchers and advocates alike, there are many reasons why the study’s findings aren’t really so bleak—and they all highlight the need to collect better data so we can understand what really works.

First, the state-run program was studied over a decade ago, and it was not particularly strong—it would not meet basic indicators of quality if evaluated today.

Second, we lack systematic data on the wait-listed children—we do not know whether these children went on to attend a different pre-K program and, if they did, how high-quality those experiences were.

Third, researchers have yet to examine how the children’s subsequent school experiences may have affected the results, a significant limitation given earlier work finding that the program’s short-term impacts varied considerably depending on the quality of the children’s elementary school teachers.

And fourth, we still don’t know how children who participated in the program are faring in later adolescence or will fare in adulthood: Other studies of pre-K attendees have detected long-term effects on outcomes like educational attainment, employment, and earnings, even after observing some short-term skill fade-out.

Only with more and better data will we be able to know where this pre-K program and others go right and where they go wrong. Better data on pre-K enrollment will help us make better comparisons.

We already know that the positive impacts of pre-K are largest when children in the comparison group have stayed at home. Many of the children who were wait-listed for Tennessee’s program likely attended other high-quality pre-K programs, such as Head Start.  

Overall, we also need better data on quality. Since 2010, there has been tremendous growth in our understanding and measurement of high-quality pre-K, and Tennessee has strengthened its program through marked investments in curricula and teacher professional development.

Over the last decade, Tennessee has also encouraged the collection of classroom quality data via the Classroom Assessment Scoring System. The state is trying to determine where further improvement is needed. Other programs, such as Head Start and initiatives in Louisiana and New York City, have made similar investments in measurement tools to strengthen data on quality. Yet many pre-K systems are still relying on quality indicators that do not factor the influence of  interactions with teachers or the use of evidence- and play-based curricula and rich instructional content.

With better data on classrooms, collected consistently across the full range of pre-K programs, policymakers will better understand the quality of existing programs and how to strengthen the components that matter most.

The measures of academic outcomes used in the Tennessee study represented a narrow band of children’s skills. They did not capture competencies like problem-solving and executive functioning—admittedly more difficult to teach and assess—that high-quality pre-K programs have been shown to enhance. There is growing evidence that such skills are more likely to be sustained as children move through elementary school.

Most data systems for early learning are not easily linked to children’s elementary school experiences and longer-term outcomes. Even in the latest Tennessee results, we don’t know how children’s experiences in elementary school and beyond relate to their pre-K experiences.

For example, the research team found that children assigned to the Tennessee pre-K program were substantially more likely to be referred to special education services starting in kindergarten. But given challenges in linking pre-K to later special education data, it is difficult to know whether this outcome was good or bad for the children. Did it help get them the services they needed, or was the referral to these services actually harmful for their learning and development?

By investing in data systems that connect early learning to K-12, policymakers will be able to better interpret study results.

Working families want high-quality care for their young children. And many states are interested in expanding their programs. Rather than curbing investments in pre-K, the Tennessee results raise important considerations about how these programs should be designed and demonstrate the importance of high-quality data for making smart policy decisions and investments that pay off for kids and families.

Meghan McCormick is a senior research associate and JoAnn Hsueh is the director of the Family Well-Being and Children’s Development policy area at MDRC.