Physical computing is a field that involves the use of sensors, actuators, and microcontrollers to build systems that interact with the environment. This approach has been used in education, particularly to teach computer science, programming, computational thinking, and design concepts. This systematic review explores the use of physical computing in K-12 education, with a focus on what educational interventions are like, how they are assessed, their effects, and how students experience them. We included 43 scientific studies that describe the methodology and assessment process used in the intervention. We found that physical computing interventions have three types of expected educational outcomes: to learn physical computing itself, to improve other learning outcomes, or to improve attitudinal variables. The interventions use a variety of educational methods to achieve their goals; as well as a variety of instruments to measure outcomes. A common lesson structure has emerged, widely adopted across classrooms, including a short lecture followed by a hands-on phase, usually including teamwork to foster collaboration. Further comparative and long-term studies are needed to understand how physical computing interventions affect students, especially those who are underrepresented.