This is a paper by Michael F. Young, Stephen Slota, Andrew B. Cutter, Gerard Jalette, Greg Mullin, Benedict Lai, Zeus Simeoni, Matthew Tran and Mariya Yukhymenko (2012) from the University of Connecticut.
Do video games show demonstrable relationships to academic achievement gains when used to support the K-12 curriculum? In a review of literature, we identified 300+ articles whose descriptions related to video games and academic achievement. We found some evidence for the effects of video games on language learning, history, and physical education (specifically exergames), but little support for the academic value of video games in science and math. We summarize the trends for each subject area and supply recommendations for the nascent field of video games research. Many educationally interesting games exist, yet evidence for their impact on student achievement is slim. We recommend separating simulations from games and refocusing the question onto the situated nature of game-player-context interactions, including meta-game social collaborative elements.
The paper can be found here.