examination of successful and unsuccessful designed movement mechanics with strategic prototyping and experimentation to isolate moment-to-moment effects of particular movement characteristics that have potential value in many interaction contexts. We are using this combined methodology in several projects in the lab, across a range of movement platforms.

September + October 2011

interactions

some designers craft very pleasurable and engaging movements, and some do not. For example, one dancing game might produce exhilaration and joy, the other tension and frustration, through the qualities of the movements that players have to perform over and over. This leads us to believe that articulating design principles and guidelines for movement feel should benefit game developers themselves, by helping them to codify and share their own successful techniques with one another.

We worked from the bottom up, looking closely at the movements, but also from the top down, looking for connections to known emotional effects of movement from psychology and communication research (see Isbister [ 8, 9] for more detail on some of the research findings we’ve found useful). We also coded the movements we observed using a subset of a taxonomy that has been used in past HCI work (for example, Sundström [ 3]), Laban Movement Analysis, which is based on dance notation developed in the early 20th century.

In parallel, we have taken up the challenge of determining how to isolate and evaluate the impact of movement on how the player is feeling. Games are complex media, with many components (animation, sound, and reward system, among others) contributing to the overall effect. We have been making use of classic social science methodology to create experiments in which we can isolate and examine the influence of movement on players.

In our first study, we had par-
ticipants play Wii games that we
precharacterized as low-, medium-,
or high-movement games [ 10].
We rigged a simple self-report
system, in which the players
to quickly rate how they were
feeling after each round, using a
two-dimensional model of emo-
tion used for quite some time in
psychology (arousal: low versus
high energy, and valence: positive
versus negative). We also video-
taped the sessions and had multiple
coders score the amount of player
movement to verify that our low/
medium/high categorizations of
the games were in fact accurate
(Figure 1 is an example of the kind
of video record we analyzed). We
found that sheer volume of move-
ment led to higher self-reported
arousal level but did not change
the valence of the players’ emo-
tional states (they reported positive
feelings in all three conditions).

Demonstrating the Value of Feel
Outside Entertainment Gaming

Recognizing that some colleagues are dubious of the value of caring about movement feel outside of pure entertainment, we have also begun to demonstrate the value of considering movement feel in productivity/non-gaming contexts. Our lab is a partner in the NYU Games for Learning Institute, which focuses on discovering and validating design patterns for building effective learning games. As part of this research, we have begun looking at whether and how the feel of movement mechanics can actually contribute to learning.

Recent research has demonstrated that striking “high power” poses for just two minutes leads to changes in peoples’ biochemistry (higher testosterone levels, lower cortisone levels) as well as greater risk tolerance when completing a sample task directly afterward [ 12]. In essence, this research supports the folk maxim that one should pose as confident in order to feel confident. What if we could leverage this fascinating phenomenon to combat math anxiety?

We have built a research prototype math game called Scoop, which allows students to practice placing fractions on a number line that they extend and contract through broad arm movements (see Figures 3a and 3b). Working in