There seems to be some papers that compare the difference between physically controlled game controllers and standard ones. The following one tackled this issue with regard to how appealin is it for the players to use these 2 kinds of interactions:
Johnson, D. M., Gardner, M. J., Wiles, J. H., Sweetser, P. M. & Hollingsworth, K. A. 2002, The inherent appeal of physically controlled peripherals, in Entertainment Computing: Technologies and Applications, eds R. Nakatsu & J. Hoshino, Kluwer Academic Publishers, USA, pp.371-378.
The author used an experimental setting to compare how players used 2 game controllers with 2 games:
Users were given the opportunity to play two Sony Playstation 2 games, SSX Snowboarding (an arcade style snowboarding game) and Time Crisis 2 (TC2; a shooting game), with both a standard controller and an alternate control device. The alternate control devices used were a Thrustmaster Freestyler Board (a snowboard shaped device which the user stands on and shifts their weight to control) for SSX, and a Namco G-Con 2 Gun (a hand held imitation pistol which is aimed at the screen) for TC2.
Then players had to use those 2 controllers with the 2 games, one after another. They used a 9 points Likert scale to evaluate the appeal of each (+ open ended questions). The results are interesting:
The qualitative findings indicate that physically implemented controllers do not always offer a greater quality of control than standard controllers, nor are they necessarily more realistic or intuitive. The findings across the two games support the first hypothesis, that the quality of control experienced by a user will be improved by physically controlled game peripherals which offer the user a more realistic or intuitive interaction.
(...)
It is important to note that while physically controlled peripherals seem to consistently lead to fun on the part of the user they do not necessarily lead to greater ease or quality of control.
There is also an account of the quantitative findings:
Overall, the ANOVA results suggest a consistent set of findings for TC2, and more complex patterns for SSX. For TC2, the physical controller (gun) was preferred over the standard controller for game performance, level of control, fun, ease of use, and overall preference. For SSX, the standard controller was generally preferred, except for ratings of fun. Also, ratings made by inexperienced and female users tended to be more equivocal than those made by experienced and male users respectively.
(...)
Overall, the results indicated that controller preferences are not simply determined by the nature of the device (standard versus physical) but instead that preferences are moderated by amount of control, intuitiveness and degree of realism experienced, and also by demographic factors of gender and experience level.
Why do I blog this? I'm interested in how people use game controllers, especially with regard to specific game play. Recently, tangible interaction is more and more trendy but there is a strong needs to evaluate how it can fit to peculiar game design. This kind of study is then very relevant for game designers to take crux decisions.