
In one way I say yes, computer games generally do follow this approach, but from another aspect I think it’s a pretty tricky accomplishment. Let me explain. The “yes” part of me thinks the act of playing computer games conforms to the constructivist approach. You become engaged in a game, providing you with an actual experience. We often play such games by choice; they internally motivate us and can provide immediate feedback which leads to self analysis. Constructivists encourage group work, that can easily be accomplished there are plenty multi-player games out there.
Now, on the other hand, can the game itself possess the constructivist qualities? This is a little tougher, although not impossible. The hard part is creating a game with the flexibility to let the player steer. This means the game must have a lot of possibilities even for a much focused topic. I would almost consider this on the way to Artificial Intelligence (AI). Where the game learns about the player and does stuff accordingly.
Fu-Hsing Tsai, Kuang-Chao Yu and Hsien-Sheng Hsiao wrote a paper and have proposed a game based learning model for designing a constructivist learning environment in online games. Their model appears valid and the explanations make sense but I feel they've missed the breadth and depth that the game maker would need to implement in order to fully follow the constructive approach. The game would have to build on the learners experiences and consider the learners point of view for subsequent levels. It is possible, but the game becomes so much more complex.
Photo courtesy of http://www.etc.edu.cn/eet/articles/learntheory/start.htm