The evidence is mounting in support of the eight-year-old "Minnesota Model" of early education investment. Now is the time to build on that model, not to abandon it.

Starting in 2008, the Minnesota Model — flexible, income-targeted, early-learning scholarships, coupled with a strong Parent Aware quality rating and improvement system to help programs adopt best practices — was first piloted by the Minnesota Early Learning Foundation. Because of very strong evaluation findings from the pilot, the approach was expanded with funding from the Obama administration's $45 million Race-to-the-Top competitive grant award, from the private sector and from bipartisan legislation. In recent years, the Dayton administration and a bipartisan group of legislators have done a good job bringing this model statewide.

According to an independent, third-party evaluation — soon to be released — this approach is working well. Minnesota children in public and private Parent Aware-rated programs based in centers, homes and schools are making significant gains on language and literacy skills, early math skills, persistence, social skills, and mental organization or "executive function" — all critically important to success in kindergarten and beyond.

Digging deeper into the evaluation, the news gets even better. Low-income children are making gains similar to the sample as a whole and actually making even stronger gains than higher-income children in executive function and language skills.

But we haven't won the battle yet. Parent Aware-rated programs are not eliminating the income-based achievement gap. But they are narrowing it, and that's encouraging.

We must do more to help low-income birth-to-5-year-olds access high-quality early-learning and home-visiting programs, because the at-risk children who are starting furthest behind often cannot catch up with only one year of help at age 4. When we start getting our youngest at-risk children into high-quality programs, we will not only narrow the achievement gap, we will prevent it.

The bottom line: The Minnesota Model is on the right track. Yet some want to head in a very different direction. Recently, Minnesota's teachers union, Education Minnesota, came out with a report advocating that the Legislature spend heavily to fund universal pre-K services for all families of 4-year-olds, including wealthier families who can already afford high-quality early education.

That well-intentioned approach is not evidence-based. Research that Rob Grunewald and I conducted at the Federal Reserve Bank of Minneapolis found that the highest taxpayer return-on-investment comes from investing in helping the most at-risk, low-income children access high-quality programs.

Other researchers have come to the same conclusion. For example, Nobel Laureate economist James Heckman, who was cited in Education Minnesota's report, has written: "The evidence in hand supports public subsidy of high quality programs targeted to disadvantaged populations." He also wrote that "there is little direct evidence on the effectiveness of the programs we study on the children of affluent families."

Spending heavily to subsidize wealthier families effectively robs Minnesota of the fiscal ability to help the most at-risk children under 5 access high-quality early-education and home-visiting services. With research showing that the achievement gap can be measured as early as age 9 months, waiting until age 4 to help the most at-risk kids is a serious mistake.

Another way the universal approach is not evidence-based is that it assumes school-based programs are the only programs capable of helping children make gains. It largely excludes other types of programs. This runs counter to evaluation findings that programs based in schools are no more effective at producing gains than highly rated programs not based in schools. The evaluation finds that all types of highly rated programs are helping kids, but also that all types of programs need more help so they can do even better.

I do concur with some arguments put forward by universal pre-K proponents. First, they note that there aren't enough scholarships available to serve all of the kids who need help. I agree. In fact, it's the single biggest early-education problem we face in Minnesota.

That's why we need to fund more scholarships, with a more flexible budget cap. Doing so would allow us to ultimately help all low-income children under 5 who are tragically being left behind. We also need more Parent Aware coaches and grants.

Second, universal pre-K proponents point out that lower-middle-class families also have needs, that these families struggle to pay for quality programs and that they have a higher likelihood of falling into the achievement gap than the wealthiest families' children.

Again, I agree. I encourage Minnesota to find a way to partly subsidize lower-middle-class families, perhaps with a sliding-scale scholarship. If Minnesota leaders first address the more urgent needs of low-income children under 5, eventually it could make sense to discuss where to draw the income lines for a partial scholarship.

But what does not make sense is moving away from the nationally acclaimed Minnesota Model of early education at a time when it is producing such impressive results. We should work to improve the Minnesota Model, not fund a less-effective one.

Art Rolnick, formerly director of research at the Federal Reserve Bank of Minneapolis, is a senior fellow at the University of Minnesota's Humphrey School of Public Affairs.