School Processes that Can Drive Scaling-Up of an Innovation, or Contribute to Its Abandonment

This paper reports findings from a multi-year study of the scale-up of Reading Apprenticeship (RA), an approach to improve academic literacy by helping teachers provide the support students need to be successful readers in the content areas. WestEd’s Strategic Literacy Initiative (SLI), began developing the program in 1995 and has since reached over 100,000 teachers in schools across the country, at the middle school, high school and college levels. In 2010, WestEd received a “Validation�? grant from the Department of Education’s Investing in Innovation Fund (i3) competition to scale-up and study the Reading Apprenticeship Improving Secondary Education (RAISE) project. For this five-year project, SLI focused on three secondary school content areas: English Language Arts, History, and Biology. From the grant, SLI not only funded an independent randomized control trial, but also a parallel effort to study the schools outside of the RCT that were to receive the same professional development and other supports in implementing RA. This scale-up study was intended to provide formative feedback to the SLI developers to help them achieve their goal to build local education agency capacity to disseminate, support, and sustain academic literacy improvement in high school subject areas within their regions. Whereas, the i3 and similar funding can “prime the pump,�? the project must build the capacity to disseminate, support, and sustain the innovation. Ultimately, adoption at the state-level, driven by local adoption and evidence of success, will keep it going. This formative evaluation of RAISE implementation collected and analyzed data on the number of trainings, the reach of the program, and the program elements that were taken up or not by participants. In this study, we followed 239 schools in four states (Utah, Michigan, Indiana, and Pennsylvania) as they participated in the expansion of the program. Schools and teachers were added to the project each year. We surveyed the teachers and principals and participated in many of the project meetings and training events. We tracked participation in the scale-up through sign-in sheets at each of the training events and maintained a spreadsheet of all the schools and teachers recording new additions each year and, less systematically, as teachers left the school or stopped participating in the program. Approximately 1720 teachers received training in the scale-up study side of the overall i3 project. To support inquiry into the scale-up process, we developed an unconventional spiraling logic model (Zacamy, Newman, Lin, & Jaciw, 2015) described below, which was inspired by the effort of putting the SLI approach to scaling up nationally together with Coburn’s (2003) insights in the processes of buy-in and commitment that make an innovation self-sustaining. The logic model pointed to activities that potentially mediated between the RAISE program and changes in educator attitudes. Our first set of studies, reviewed below (and reported in Zacamy, Jaciw, Lin, & Newman, 2014; Zacamy & Newman, 2014; Zacamy et al., 2015), examined survey responses to gauge adoption of RA, buy in by teachers and school administrators along with measures of participation in program activities. These analyses provided useful insights and suggestions about malleable factors that could be focus of future improvements. The surveys, however, did not directly address the actual scaling-up of RA, that is, the numerical increase in number of teachers and schools using RA. This is an issue addressed in our logic model, which hypothesized that “as capacity and support builds, we expect districts and schools to increase the numbers of teachers implementing RA; that is, schools will send more teachers to RAISE training and spread the RA ideas to other districts and schools.�? We were also sympathetic to the idea that scale-up can be measured in terms of the increasing numbers of participants over time (Slavin, 2002). With these considerations in mind, we harvested the data from the participant tracker and created a dataset of schools and teachers with which we could measure the growth and loss of RAISE participation within states, districts, and schools. These increases or decreases in number of participants becomes an outcome that can be considered the goal of the investments in professional development that are hypothesized to promote intermediate outcomes such as local level ownership and commitment. The current paper reports our analysis linking growth in numbers of participants to contextual variables, e.g., school size, percent free and reduced price lunch (FRPL), to program events, e.g., monthly meetings, and to teacher and principal level of commitment.