So today I submitted my thesis. It spans 395 pages, with over 40 pages of references & 3 pages of acknowledgements. I've not been very vocal about my research on Twitter in the past 3years, preferring to Tweet memes, so here's a summary! Thanks HPSC research degrees for the star! 
I started with the REALLY broad research question: "Can videogame avatars affect health-related behaviours?" This narrowed down to "Can the physique of an avatar affect exercise intensity?" which led me down unexpected paths! 1/n
I started with a pre-registered systematic review of the existing literature. This was meant to be a meta-analysis, but I could only find 14 relevant studies! My conclusion was: the area is sparse, varied, & (surprise!), likely biased! 2/n
One recurrent finding (N=3) was that fat avatars make people move slower when playing the Nintendo Wii, as players temporarily assume the identity, & adopt the stereotypical behaviours of the avatar. This is called the #ProteusEffect. 3/n
This led some authors to suggest limiting avatar body types to thin or athletic! 4/n
I interviewed 20 people about gaming & avatar design. They played an Xbox Kinect climbing game that tailored an avatar to look like them, but with an athletic physique. People preferred to choose their own ideals, & were critical of the enforced idealisation 5/n
This led me to study how athletic and fat people are stereotyped. STs are the driving force behind the Proteus Effect. I designed a set of avatars to represent these groups & used word applicability rating scales to explore stereotypes. 6/n
The stereotypes were complex and ambivalent. Using parallel analysis I found three ST factors: Determination, Arrogance, and Sociability. Plus sized people were rated as less arrogant and determined, & more friendly than athletic people 7/n
This was observed both when participants responded to both verbal descriptors of groups, and to the set of avatars that I designed. With evidence that my avatars invoked these STs, I looked into how to break these, using science! 8/n (IMG credit XKCD)
I introduced counter-stereotypical behavioural information by animating my avatars to run on the spot. Lo and behold, the applicability of weight-stigma related words (e.g. lazy) was reduced for fat avatars! 9/n
My reasoning was that if the Proteus Effect holds, reducing the perception of weight stigma should improve movement during an exercise videogame. So I built a running exergame that had double blind random assignment to my avatars! 10/n
In the 8 weeks that I actively collected data, I could only find 65 people, & some of these were excluded due to software errors. This was too small to draw firm inferences. I will be collecting more data in the new year (it's a Bayesian sampling plan) 11/n
So far though, people in the large bodied condition were on average faster than those in the athletic condition - so the opposite of what was predicted. I have huge error bars, so take with a pinch of salt! 12/n
As a side hustle, I also wanted to see if using an athletic avatar affected peoples confidence in their ability. I used a base-rate neglect paradigm, & found that people with athletic avatars paid less attention to statistical information 12/n
I intended for this to be a proxy for arrogance; & I'd be interested in hearing peoples thoughts on this. It may apply to exergames- if a player has an athletic avatar they might go straight to hard mode, instead of practicing on easy! 13/n
My experimental data were REALLY dependent on priors, & I could actually flip the observed effect! Fortunately this one, & most of my other studies are pre-registered, or the temptation of a quick finish may have got the better of me! 14/n
My take-home message is that there may be other ways to use avatars to improve exercise during an exergame; banning physiques may be extreme and unwanted! 15/n
I will be making all data & scripts, registrations, etc available on OSF in the near future (just let me get through my viva first!). I'd like to thank my supervisors @Dr_JCole Sarah Grogan and @Niki_J_Ray @Psypag and @MMUPsychology for various grants for data collection! n=16
