Julia Isaacs of the Brookings Institution has a just-released paper (March 19, 2012) that provides valuable comparisons of preschool versus other methods of increasing school readiness, at kindergarten entrance, for children from low-income families.
The backdrop to this paper is the finding, based on a previous paper by Isaacs and University of Wisconsin Professor Katherine Magnuson, that the “school readiness gap” between children from families below the poverty line, versus families with incomes above 185% of the poverty line (the cutoff in American schools for eligibility for a subsidized school lunch), is 27 percentage points. Children from poverty families have a school readiness percentage at kindergarten entrance of 48%, versus 75% for children from families above 185% of the poverty line.
Isaacs in this paper, and in the preceding paper with Magnuson, looks at how school readiness at kindergarten entrance is related to preschool attendance, as well as other variables such as maternal smoking, and family income. In the current paper, Isaacs uses these estimates, along with estimates of program costs and program effectiveness, to estimate the effectiveness of various interventions to improve school readiness. She asks two useful questions about effectiveness:
(1) How effective in raising school readiness is the program or intervention per dollar of program cost?
(2) Given what we know about how many children can reasonably be expected to be affected by this intervention when expanded to full scale, how large an effect might the intervention be expected to have in reducing the school readiness gap for all poor children?
Isaacs makes the following two conservative assumptions about the effectiveness and plausible scale of the preschool program.
(1) She assumes that a preschool program will need to cost $7,200 to raise kindergarten readiness by about 9%, which is about one-third of the total gap between low-income and moderate income children in school readiness.
(2) She assumes that the scale of her assumed preschool program is limited to increasing preschool enrollment among low-income children from current levels to 66% to enrollment levels of 90%.
The first assumption is conservative because there are studies that estimate that cheaper preschool programs than assumed by Isaacs can raise school readiness by about the same amount as Isaacs assumes, and that slightly more expensive preschool programs can raise kindergarten readiness by twice as much. For example, in my recent study with Bill Gormley and Shirley Adelstein of Tulsa’s universal pre-K program, we found that a half-day pre-K program for 4-year olds that cost about $4500 per child could raise the test score percentiles of children from poor families by about 12 percentiles, which implies that it would probably raise kindergarten readiness by close to that amount. A full-day pre-K program that cost about $9,000 per child could raise test score percentiles by 18 percentiles, which would imply a similar increase in kindergarten readiness.
The second assumption about scale of pre-K reform is conservative because it assumes that the only impact of moving to universal pre-K is to increase pre-K access for children from poor families from 66% to 90%. But instituting universal pre-K for all low-income children would probably also improve the quality of pre-K for the 66% who are already in some type of program that is called preschool.
Even with these two conservative assumptions, she gets the result that moving to universal pre-K for poor children would have the potential for improving school readiness of ALL poor children by 2.1% (e.g., a 9% improvement for the 24% of children who are newly added to preschool). This policy intervention would close one-tenth of the school readiness gap between poor children and moderate income children, at a nationwide cost of only $1.6 billion. This is quite a powerful effect for such a relatively cheap intervention. And I think it would be easily possible with other defensible assumptions to get effects on school readiness of 3 or 4 times as much for a program of scale expansion and quality improvement that might cost between $2 and $3 billion nationwide. Both cost-effectiveness and potential scale of preschool reform may significantly exceed Isaacs’s estimates.
Even with these conservative assumptions, preschool compares quite favorably with other interventions when judged by its effects on school readiness at kindergarten entrance. This shouldn’t be a big shock, as kindergarten readiness is one of the key outcomes that high-quality preschool should be designed to significantly affect.
For example, smoking cessation programs for low-income mothers have a similar cost-effectiveness to preschool in raising school readiness, because these programs are so cheap per mother treated. But these programs have only one-20th the potential overall effect on school readiness for low-income children when expanded to full scale, primarily because (1) only 20% of low-income mothers smoke, and (2) smoking cessation programs only reduce smoking rates by 6% of those treated, both of which factors limit the potential scale of aggregate effects of these programs.
Nurse home visiting programs for disadvantaged first-time mothers also improve school readiness. However, nurse home visitation is about twice as expensive per a given improvement in school readiness as preschool, and only has about one-fourth the potential scale in its aggregate effects on school readiness. This is primarily because nurse home visitation programs have broader goals than school readiness for children, because such programs are limited in their coverage to first-time mothers, and because these programs show the greatest effects for the low-income mothers at highest risk.
Finally, we can improve school readiness by simply giving low-income families more money. While income transfers may have many broad social benefits for low-income families, Isaacs’s estimates imply that such income transfers would have a cost, for a given increase in school readiness, of about 10 times the cost of achieving the same school readiness gains by expanded preschool. Again, this should be no surprise: simply providing income transfers is not as targeted in its benefits for children in terms of school readiness as is true of providing preschool.
Therefore, Isaacs’s estimates, which are consistent with other estimates, suggest that moving to universal preschool for poor children is a cost-effective way of improving their school readiness that can be expanded to have sizable aggregate effects.
But is school readiness important? My prior paper with Gormley and Adelstein, which is based in turn on prior work by Harvard Professor Raj Chetty, suggests that how students are doing as of kindergarten is important to adult outcomes. Improving a child’s kindergarten test scores by 1 percentile is estimated to raise average adult earnings by about one-half of 1%. A 1 percentile increase in a group of children’s test school performance is likely to translate into an increase in kindergarten readiness by a similar amount. Therefore, if we can increase the kindergarten readiness of all poor children by 10% or so, we would be likely to improve their average lifetime incomes by at least 5%. This would be a quite large dollar figure.
In sum, expanding quality preschool to all low-income children can plausibly have substantial effects on improving the U.S. income distribution, as I have argued before. Furthermore, universal preschool for children from working class and middle class families will also have sizeable effects on the future income of the broad middle of the U.S. income distribution.