Even though virtual labs help students learn science content, little is known about how well students can later apply this learning to other contexts or tasks when compared to students who performed physical labs. The goal of this study was to understand how students who perform physical versus virtual labs were able to later apply what they learn to a new context and a more intricate physical lab. We also explored whether reducing the complexity of the physical lab, by setting up the apparatus before students conducted experiments, supported students' learning of physics concepts and relationships when compared to students who performed a virtual lab. Using a quasi-experimental research design, we randomly assigned 26, seventh and eighth grade classes from seven teachers' classes into two conditions, the virtual or pre-set-up physical pulley lab conditions. Using data collected from 385 students and 188 groups, we found that students who were in the virtual lab condition learned significantly more about the mechanics of pulleys than students in the physical lab condition as assessed by a content knowledge pre to posttest. We also found that students in the virtual condition performed better on answering and explaining a real-world scenario application question and took no longer to set up a more complex pulley system in small groups than students in the physical condition. We discuss limitations, implications, and directions for future research.