Many find statistics confusing, and perhaps more so given recent publicity of problems with traditional p-values and alternative statistical techniques including confidence intervals and Bayesian statistics. This course aims to help attendees navigate this morass: to understand the debates and more importantly make appropriate choices when designing and analysing experiments, empirical studies and other forms of quantitative data.
Making Sense of Statistics in HCI: From P to Bayes and Beyond – introduction
1. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
Making Sense of Statistics in HCI:
From P to Bayes and Beyond
Alan Dix
http://alandix.com/statistics/
2. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
[-0.1,+3.2]
[ 0.2, 3.7 ]
95% conf. int.
p-values, confidence intervals, Bayesian stats
xxx
what does it all mean?
3. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
confused?
4. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
focus on understanding
concepts and ideas
5. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
make the most of your
empirical effort and
avoid misleading results
6. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
why is probability so hard?
subconscious behavioural conditioning
needs lots and lots of exposures
associative and semi-probabilistic
conscious thinking and learning
learn from single example
single model of the world
… but probability hard
* see also: http://gladwell.com/blink/the-second-mind/
https://archive.org/details/controllingbehaviorthroughreinforcement
7. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
why is statistics so hard?
some maths
need to understand both
real world
8. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
the ‘real’ world
the sample
actual measured data
the population
large set from which the data is drawn
especially for surveys etc.
the ideal
the ‘typical’ user, the fair coin
unrepeatable events – the fall of a raindrop
a theoretical distribution
9. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
the job of statistics
10. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
overview – four parts
wild and wide
exploring randomness, uncertainty and 'distributions’
doing it
alternative statistical analyses: the ubiquitous 'p' to Bayesian
gaining power
avoid the dreaded 'too few participants’
so what?
making sense of your data and avoiding the pitfalls
11. Making Sense of Statistics in HCI: From P to Bayes and Beyond – Alan Dix
more …
various things including Javascript demos at:
http://alandix.com/statistics/
any updates of these materials, or further information
for CHI course attendees at:
http://alandix.com/statistics/course/chi2017/