Using Iterative Experimentation to Accelerate Program Improvement: A Case Example

This article offers a case example of how experimental evaluation methods can be coupled with principles of design-based implementation research (DBIR), improvement science (IS), and rapid-cycle evaluation (RCE) methods to provide relatively quick, low-cost, credible assessments of strategies designed to improve programs, policies, or practices. This article demonstrates the feasibility and benefits of blending DBIR, IS, and RCE practices with embedded randomized controlled trials (RCTs) to improve the pace and efficiency of program improvement. The illustrative case is a two-cycle experimental test of staff-designed strategies for improving a workforce development program. Youth enrolled in Year Up's Professional Training Corps (PTC) programs were randomly assigned to "improvement strategies" designed to boost academic success and persistence through the 6-month learning and development (L&D) phase of the program, when participants spend most of their program-related time in courses offered by partner colleges. The study sample includes 317 youth from three PTC program sites. The primary outcome measures are completion of the program's L&D phase and continued college enrollment beyond the L&D phase. The improvement strategies designed and tested during the study increased program retention through L&D by nearly 10 percentage points and increased college persistence following L&D by 13 percentage points. Blending DBIR, IS, and RCE principles with a multi-cycle RCT generated highly credible estimates of the efficacy of the tested improvement strategies within a relatively short period of time (18 months) at modest cost and with reportedly low burden for program staff.