Researchers in Thailand did a preliminary study on this; see Teerawat Kamnardsiri et al: "The Effectiveness of the Game-Based Learning System for the Improvement of American Sign Language using Kinect", Electronic Journal of e-Learning 15(4): 2017-283, August 2017.
The researchers divided 31 deaf learners (average age: 21) into two groups: a group of 17 learners who used a Kinect game, and a control group of 14 learners who were taught sign language in a traditional way. Prior to the experiment, there was a test of the learners' sign language skills, which showed no significant differences. After the experiment, there was a new test, which showed statistically significant differences between the two groups; the "Kinect" group performed better than the control group.
There are a few limitations to the study:
- The test involved a very small set of words or concepts, i.e. a much smaller set than what deaf learners need to acquire.
- The sample of learners was very small.
- The experiment duration (six months) was considered rather short.