Interactive cognitive complexity theory suggests that simulation games are more effective than other instructional methods because they simultaneously engage trainees’ affective and cognitive processes (Tennyson & Jorczak, 2008). Meta‐analytic techniques were used to examine the instructional effectiveness of computer‐based simulation games relative to a comparison group (k= 65, N= 6,476). Consistent with theory, posttraining self‐efficacy was 20% higher, declarative knowledge was 11% higher, procedural knowledge was 14% higher, and retention was 9% higher for trainees taught with simulation games, relative to a comparison group. However, the results provide strong evidence of publication bias in simulation games research. Characteristics of simulation games and the instructional context also moderated the effectiveness of simulation games. Trainees learned more, relative to a comparison group, when simulation games conveyed course material actively rather than passively, trainees could access the simulation game as many times as desired, and the simulation game was a supplement to other instructional methods rather than stand‐alone instruction. However, trainees learned less from simulation games than comparison instructional methods when the instruction the comparison group received as a substitute for the simulation game actively engaged them in the learning experience.