A Half-baked Cake: Idealised Avatars In An Exergame


The Avatar

In August of last year I wrote a blog advertising my interview study on experiences of play an exercise game with an avatar that looks similar to the player.

Last week I presented some of my findings at a Games Research Network seminar and thought I would summarise the talk in this short blog post

I started the talk by defining the term avatar which has only recently received a thorough explication by Nowak and Fox (2018). In short, it is the representation of a person in digital space. This is not restricted to visual representations, but can be text, sound, and in some cases haptic. For my purposes, I am looking at avatars that specifically look like a person, and in particular, videogame characters.

My area of interest is the way in which avatars may subconsciously persuade individuals to behave differently. Much of the research in this area has stemmed from a study by Yee, Bailenson, and Ducheneaut (2009) who report a tendency for people to behave differently when embodying different avatars in an immersive virtual environment. This effect was named the Proteus Effect after a Greek god who could assume many different forms. For instance, participants who observed an attractive avatar when looking in a virtual mirror were more affable and walked closer to a confederates avatar then those who observed an unattractive avatar.

Where the Proteus Effect relies on the embodiment of an avatar for the behaviour change to occur, simple observation of a model has also been shown to affect behaviours when the model looks similar to the individual. Fox and Bailenson (2009) found that participants who observed a virtual self representation losing weight whilst exercising and gaining weight when stationary performed more voluntary exercise repetitions than those who observed no feedback. This is explained by Social Cognitive Theory whereby individuals are alleged to imitate behaviours that are rewarded through a process called vicarious reinforcement (Bandura, Ross, and Ross 1963).

These theories have been applied to various videogame and studies on videogames. You may remember the Wii Fit function whereby if a player did not exercise their Mii would gain weight. This was also found in Grand Theft Auto San Andreas.

My interview project aimed to investigate how participants responded to an athletic avatar that was tailored to look like themselves through a camera process. The intention was to take a critical and qualitative view of the previous studies I have mentioned and find out what participants thought about being represented as a heroic athlete.

I interviewed 20 people (9 males, 11 females) about their gaming habits and health attitudes. Participants then generated an avatar using an automated process before playing two games of Kinect Sports Rivals (Rare 2014). I then asked them about the general experience and how they believed it could change peoples health behaviours. I ended up with 30 hours of interview transcripts which I analysed using thematic analysis and I will present a few of the themes from the dataset below.

Response to the Avatar

Some participants placed a lot of trust in the game’s ability to portray them, with one participant stating ‘I expect that camera to know what it’s doing’. This was an interesting finding because it was driven by the tailoring process. When people generate avatars their attention is drawn towards a camera which are used to reflect reality rather than the videogame console and software which are designed to remove individuals from reality. This trust can be beneficial or demotivating depending on how the player feels about the avatar they are presented with.

Other participants were happy with the increased athleticism of the avatar and saw it as flattery, for example:

I feel like it’s a lot portrayed me a lot better than I actually am.

In some cases this was also reflected by their own avatar creation behaviours with some participants claiming to create characters in games like Skyrim that reflected their ideal self.

Other participants rejected the idealised avatar as being similar to themselves and found it to be overly sexulised and idealised as reflected by one female participant: > ‘I would lump this together in with all the other sexist shit that shouldn’t be’

Other participants said that they did not identify with the avatar because it did not reflect their changeable identity. One fashion conscious participant said of the avatar’s outfit:

I’d never wear anything like that.

Other sticking points included hair and make-up such that the character on screen did not represent their fashion sense.

Response to the Game

I also asked participants about their experience of playing the game itself. Responses were mixed both between and within participants!

The majority of participants initially claimed that the game was ‘quite a fun experience’, suggesting that it would make a good party game. Some also though it would be a good way for people who do not exercise to get a foot in the door with physical activity.

However, other participants saw it as a ‘tech demo’ or a something that was neither a good game nor a good avatar experience, with one participant saying:

It seemed to be a bit of a like half-baked cake.

The largest complaint about the game was the controls. The kinect camera was often unresponsive, laggy, or in some cases seemed to take control away from the participant. This is reflected in the following statement from in the game:

Right Hang on, just go up now, f*%k, no way, go go go, jump jump jump, oh no no no!

In my discussion I suggested that KSR has its own rhetoric within the avatar generation system. Not only are characters thin and muscular, but they are unable to be anything but this. There is no room for participants to actually put themselves as they see themselves in game. This was a problem for some participants whose ideal body image did not fit in with athletic ideals. There are also no options for plus size characters fit and athletic. For some people this was fun and they enjoyed the athletic avatar but for others is led to a self discrepancy which can be demotivating.

My upcoming research projects will involve revisiting some of the earlier studies to explore integrating body positive messages in exercise games rather than locking people to a social ideal.


Thanks to my supervisors Dr Jenny Cole, Professor Sarah Grogan and Dr Kevin Tan for their input and guidance on the on the designa and analysis of the study.


Bandura, Albert, Dorothea Ross, and Sheila A Ross. 1963. “Vicarious Reinforcement and Imitative Learning.” The Journal of Abnormal and Social Psychology 67 (6). American Psychological Association: 601.

Fox, Jesse, and Jeremy N Bailenson. 2009. “Virtual Self-Modeling: The Effects of Vicarious Reinforcement and Identification on Exercise Behaviors.” Media Psychology 12 (1). Taylor & Francis: 1–25.

Nowak, Kristine L, and Jesse Fox. 2018. “Avatars and Computer-Mediated Communication: A Review of the Definitions, Uses, and Effects of Digital Representations.” Review of Communication Research 6: 30–53.

Rare. 2014. “Kinect Sports Rivals.” Redmond, Washington, USA: Microsoft Studios.

Yee, Nick, Jeremy N Bailenson, and Nicolas Ducheneaut. 2009. “The Proteus Effect: Implications of Transformed Digital Self-Representation on Online and Offline Behavior.” Communication Research 36 (2). Sage Publications Sage CA: Los Angeles, CA: 285–312.