This seminar took place on June 22, 2017 @ 1:30PM-4:30PM (UTC+2). The link to the video is below.
There has been two talks, the first one by Pierre Dragicevic and the second one by Yvonne Jansen. Only the video of the first one is available at the moment but the webpage will be eventually updated.
Here is the link to the video. If you have any trouble with the flashplayer, here is a direct link to the mpeg4 video. Here are additional references used by Pierre during his talk.
To interact during the presentation (questions, comments), we used this pad that we will keep for the record.
Statistical dances: why no statistical analysis is reliable and what to do about it (Pierre Dragicevic)
We need to improve the way we do statistics in HCI, but more training in statistical theory is not enough. We also need good “intuition pumps” to develop our statistical thinking skills. In this talk I explore the basic concept of statistical dance. The dance analogy has been used by Geoff Cumming to describe the variability of p-values across hypothetical replications. Through visual examples, I show why any statistical analysis and any statistical chart actually dances across replications. I discuss why most attempts at stabilizing statistical dances are either insufficient or misguided. The solution is to embrace the uncertainty and messiness in our data. We need to develop a good intuition of this uncertainty and communicate it faithfully to our peers. I give a few tips for conveying and interpreting interval estimates in our papers in a honest and truthful way.
Pierre Dragicevic is working in the Aviz team at Inria as a permanent research scientist (CR). He studies information visualization and human-computer interaction. He co-signed many research articles with p-values until he grew dissatisfied and banished p-values from all his publications. Since then, he has been promoting an approach to statistics based on planned analyses, interval estimation, graphical communication, and nuanced interpretations.
In 2010 Dana Carney, Amy Cuddy, and Andy Yap published a study on the effects of “power poses” on hormone levels and risk-taking behavior. Their findings were widely publicized through Amy Cuddy’s TED talk and have inspired much work building on these findings. Yet recent developments suggest that the effect might not actually exist and that instead many failed replications fell victim to a publication bias. In this talk I will show how we ourselves attempted to apply power poses to interface designs, how we failed to replicate effects of power poses, and the difficulties we encountered in publishing our negative findings. I conclude by discussing the value of sharing negative findings with the research community and how it conflicts with our current publication system.
Yvonne Jansen is a CNRS research scientist at the ISIR lab at UPMC working within information visualization and human-computer interaction. Her research focuses mostly on interaction with physical interfaces or spaces. She is also interested in research methodology, effective communication of statistics, and reproducible science in general.
- Grenoble: François Bérard, Renaud Blanch, Arnaud Legrand
- Grenoble: Imag Amphitheater on Campus
None