Association for Public Policy Analysis & Management

Successfully scaling-up education programs is a long-standing problem and one that is currently particularly pressing in preschool, as cities and districts around the country work to rapidly expand their public preschool programs.  The Boston Public Schools has had marked success in scaling its prekindergarten model (called K1). The program has the highest level of instructional quality of any evaluated large-scale prekindergarten in the U.S. (Weiland, Ulvestad, Sachs, & Yoshikawa, 2013) and has substantial impacts on children’s language, literacy, mathematics, and executive function skills at kindergarten entry (Weiland & Yoshikawa, 2013).

However, as the City of Boston pursues a mixed-auspice approach to public prekindergarten, the district is facing a new challenge: scaling its model outside the public schools, to structurally weaker community-based preschools. At APPAM, we will present final results from Boston’s K1 initial scale-out effort: a multi-year pilot program called K1 in Diverse Settings (K1DS). Over 2.5 years, this pilot introduced the hallmarks of the BPS K1 model—developmentally appropriate, research-based language, literacy, and mathematics curricula supported by regular one-on-one coaching and training—to 14 classrooms in 10 community-based organizations (CBOs) which offered education and care services to 4-year-olds in high-poverty areas of Boston, MA.

We used a mixed methods approach to evaluate the K1DS pilot project. Data on language, literacy, and mathematics instructional quality were collected in each classroom at baseline before the intervention began (January 2013) and at the end of each school year (May 2014 and May 2015). We also conducted interviews with all key stakeholders at the beginning and end of each school year.

We found that sizable gaps in classroom quality between CBO K1DS and BPS K1 classrooms shrank in the first 1.5 years of the intervention, but initial progress was fragile. Figure 1 displays K1DS classroom means for the literacy/language and mathematics quality measures across the three time points. For all measures, there was an increase in quality from baseline to spring 2014 and a decline in quality from spring 2014 to spring 2015. Notably, spring 2015 means are modestly higher than baseline means for three out of the four subscales, meaning some of the initial improvements in literacy/language and mathematics instructional were sustained, but that quality ultimately fell short of that of BPS K1 classrooms. At APPAM, we will expand upon these findings.

We also found that CBO K1DS classrooms varied in fidelity to curricula. Four of the 14 classrooms had low fidelity for both curricula; 7 reached a moderate level; and 3 a high level. At APPAM, we will expand upon this finding by examining the relationship between fidelity and classroom quality.

Finally, we identified important structural barriers which help contextualize why some classrooms were able to implement the intervention with fidelity while others were not. At APPAM, we will describe these barriers in detail and discuss policy implications that stem from this work. In particular, we will focus on the structural supports that might be leveraged to ensure successful scale-out and scale-up of pre-k programs nationally.