Recently, we performed an experiment dressed up as a gameshow with the title Lies, Damned Lies and Biodata (a nod to “Lies, damned lies and statistics” – Benjamin Disraeli for those who failed to pick up the reference). In practice we aimed to explore to what extent lay-people could “read” biodata when presented to them in a fairly raw form (just some line graphs and an EEG heat map).
The aim of the game was for the audience to guess whether or not a person was lying or telling the truth by looking at the presented biodata. The liar (or not) stood at the front of the stage and a picture was shown. They were asked a question about the picture and secretly informed whether they should lie or tell the truth. So for example the picture might be of their house – and they might be asked “is this your house” and told to lie, alternatively they might be asked to tell the truth. Or they might be shown a completely different house and asked the same question with the same conditions. It was a little confusing, but ultimately quite a successful experiment.
We dressed it up as a gameshow, complete with prizes, cheesy music and all the verbal silliness implied in the title. On the whole, the day seemed to go down fairly well. All the players did pretty much as was asked, with a couple of accidental exceptions and the audience, few as they were in the end did kind of get behind the game.
I take from this experience several key findings:
- Biodata is difficult to read without some explanation
- Playing with truth and lies is fun as long as everybody gets in on the game
- I’m not a natural born presenter
We hope to explore for the recorded data what subsets of information are usable to increase or decrease the facility we have for lie detection. Obviously a standard lie detector uses many of these biosensors, so there is a precedent there – my interest is in whether there is a way to present this stuff that allows it to be read without any special training.
All in all a fun day out.