Research in Focus: Using data to identify unconscious bias
'Pace and Power: Removing unconscious bias from soccer broadcasts', by Sam Gregory, Devin Pleuler, Daniel Daly-Grafstein, Yang Liu, and Paul Marchwica
[Link to the NESSIS presentation here]
Why it's worth your time
Unconscious bias is important to address but, to address it, it needs to be identified. Using tracking data and computer animation to anonymise data and make it look like watching football on video, this gets easier to do.
What it says
The topline: There was a statistically significant difference in which team, in a Poland vs Senegal game, was judged as 'more athletic', depending on whether respondents watched a normal video or an anonymised, data-derived render. The video-watching group tilted towards Senegal; the render-watching group towards Poland.
As you may have gathered, the study design split respondents into two groups, one of which got a normal video to watch and the other a computer-generated render, before being asked various questions comparing teams in the clips.
The study also asked respondents to watch video from a men's English League 2 match and women's NWSL match to assess gender bias in judging quality of play. In that, the group who watched the video had a higher rate of choosing the men's match compared to the group who watched the render, but not at a level that met statistical significance in this study.
What's cool about it
The approach is pretty ground-breaking: until this type of technology and data was available it wasn't possible to 'anonymise' watching football to check for this kind of bias. While the researchers note that this study was relatively small (105 participants) and used relatively short clips, it lays a framework for future research. With advances in rendering and increasing availability of 3D skeletal data, we could see a lot more studies like this in the future.
Maybe this approach could also find a use within clubs, national football associations, or training courses to train scouts and analysts.
[Link to the NESSIS presentation here]
'Research in Focus' is like SparkNotes for football analytics: summarising and analysing the best research out there. Get Goalside supporters get access to posts a month early. Follow this link for the list of all Research in Focus pieces.