Q&A with Anamarie Whitaker on Pre-K Effectiveness

What prompted you and your colleagues to publish this working paper at this specific point in time?

We have been reviewing the evidence on pre-K programs and noticed puzzling trends in more recent program evaluations—specifically from studies using rigorous methods (e.g., lotteries or randomized control trials). This is notable because there are few evaluations of modern programs using such designs – five to our knowledge of programs in: Tennessee, Boston (2), North Carolina, and Georgia.

On average, results suggest that the programs located in these states produce an immediate positive impact on children’s academic development. By the end of kindergarten, the positive effects mostly fade and in two of the three evaluations of at-scale state pre-K programs (Tennessee and Georgia), there are negative, significant impacts on children’s achievement in later elementary school compared with children that did not attend the pre-K program.

So, these are puzzling, particularly when compared with the more uniformly positive results from the Perry Preschool and Abecedarian evaluations. There are a lot of unanswered questions about why programs may be producing negative results and about positive long-term but not intermediate results from the program evaluation in Boston.

Our group wanted to explore commonly discussed theories for why recent results differ from those of earlier programs. This includes program quality, changes in conditions and school environments from the 1960s to 2019, and changes in pre-K instructional practices. Specifically, we wanted to determine which theories could be supported with available evidence.

In the paper you compare the cost of the Perry Preschool and Abecedarian pre-K programs of the 1960s and 1970s with the funding levels of more modern pre-K programs. What do those differences look like in terms of 2023 dollars and why do funding levels matter?

In the paper, we include the cost of modern day high-quality, full-day programs, which range from approximately $15,000-16,500 per child per year, compared with the cost of Perry Preschool and Abecedarian, which were more than $40,000 per child (average 1-2-year cost) and approximately $120,000 per child (5-year cost), respectively. All the figures are in 2023 dollars, so they are apples to apples comparison in terms of present-day costs.

Funding for programs matters if we believe that funding is related to higher quality programming. For instance, more funding could translate into higher educator wages and benefits, more classroom materials, and improved teacher-child ratios. Our programs today are operating on a much larger scale than Perry or Abecedarian, so it is not surprising that funding levels differ. However, the discrepancies between modern day and demonstration programs raises a larger question: How much would we need to spend per child, and what would we need to spend the money on (e.g., educator salary) to realize the benefits produced by earlier programs?

In the paper, you note that, “Societal changes in the last 50-60 years now provide all children with some of the advantages that the early preschool projects provided in the 1960s and 70s.” What are some examples of those positive societal changes?

Funding for economic supports, such as Temporary Aid for Needy Families (TANF) and the Supplemental Nutrition Assistance Program (SNAP), and per child spending at the federal level has increased quite substantially from the 1960s. We also see declines in infant and child mortality, suggesting improved conditions due to either environmental and/or medical advancements. Earlier preschool programs, particularly Head Start, provided families access to services, specifically medical and dental care and referrals, that are now more widespread through other social programs such as Medicaid and the Woman, Infants and Children (WIC) program. K-12 schools also spend about four times the amount per student, inflation-adjusted, compared with 1960. Thus, the earlier preschool programs were providing services that, while still may not be universal today, are more available than before.

So, in a study where some children randomly get a preschool slot and others don’t, the children who don’t will have access to more, and plausibly higher-quality services, than they would have had the study been done 50-60 years ago. Of course, this is mostly good news! But it is also potentially a reason why we are not seeing the same types of impacts found for those initial programs.

You also point out that, “changes in public education could account for why impacts of current programs may diminish more rapidly than for earlier programs.” What are some examples of those changes in public education, both in pre-K and kindergarten?

We believe this is an important area for future research. Evidence does suggest that kindergarten and pre-K are both becoming more academic (e.g., spending more time especially on literacy instruction). Over the past twenty-five years, instruction switched from child-centered activity centers that focused on teaching social skills and general knowledge to intentional instruction that increasingly focused on teaching early literacy skills and, to a lesser extent, early math skills. This academic instruction often requires preschoolers to sit quietly in large groups while the teacher leads the lesson. Kindergarten teachers also are expected to teach the same early academic skills under their state’s learning standards, so many pre-K children are taught the same skills again in kindergarten.

In addition, the increased time spent on academic skills, especially early literacy skills, replaced some of the time that was spent in earlier preschool programs on teaching language, general knowledge, and social skills. It was these skills that largely accounted for the long-term impacts on adult outcomes in the early preschool projects. For these reasons, we believe more research is needed to really understand how classroom instruction—more the mode of instruction and the content—in pre-K and kindergarten is related to medium or long-term outcomes and how best to align instruction throughout the early grades.

The paper ends by stressing the importance of striving to make modern pre-K programs as effective as possible along with recommendations regarding the direction future pre-K research should take to accomplish this goal. What are some of those recommendations?

We really need to understand the mechanisms behind how preschool can benefit children and we need high-quality longitudinal research on programs. For instance, we only have one randomized evaluation from a modern day program that follows children into high school. More rigorous research (e.g., randomized evaluations) can help us better understand what the causal impacts of pre-K participation are on children’s short and long-term outcomes—really what is realistic to expect from preschool programs. Right now, we are continuing to rely on the demonstration programs from the 1960s and 70s for this evidence. We now have evidence from three randomized studies that pre-K impacts produce immediate positive effects on children’s achievement. But then the effects fade and, in two of those studies, negative academic impacts are found in elementary school.

We must take these findings seriously and seek to understand why this is happening. In the long-term, we think that investments in research to follow the children from all of these studies into adulthood will help us piece together a puzzle of how short-, medium-, and long-term effects are related.