As host of the UK version Chris Tarrant would say, ''this is serious business''. A previous post of mine made a light-hearted comparison of the show to investing. Ha-ha, the joke is on me. I've just discovered a heavy duty study on the subject. Do you like equations with lots of Greek letters, then download Who Really Wants To Be a Millionaire: Estimates of Risk Aversion from Game Show Data, a paper published in 2005 by Roger Hartley, Gauthier Lanot and Ian Walker of Warwick University? Attached is a sample ... just don't ask me to explain.
The paper includes some fascinating trivia based on a complete inventory of the first 11 series of UK shows up to June 2003 (the authors had the cooperation of the show sponsors):
- 3 out of 515 contestants won the £1 million top prize (using my higher math skills, that works out to 0.6% not the 1-2% I from my first source), which is the same number who went away with nothing
- the mean of winnings was much higher at £54k than the £16k I'd guessed, though the standard deviation was a whopping £106k (i.e. those infrequent big prizes distort the stat)
- 2/3 of people voluntarily stopped by deciding not to answer versus the 1/3 who had to stop by getting a question wrong
- the quitters left with an average of £72k while the wrong guessers ended up with an average £17k (and this latter figure includes the £1 million winners who the authors decided had not quit voluntarily!); unfortunately the authors don't comment on this difference and I wish they had ... maybe it's called knowing when to quit while you're ahead?
- the probability that a phone-a-friend will know the answer is only about 40%
- the ask-the-audience lifeline is as valuable as both the phone-a-friend and the 50-50 put together; is this a manifestation of the wisdom of crowds / markets?
- 3/4 of the contestants were men, who the authors say are less risk averse than women
- the early round questions are more weighted with pop culture and sport
- the show's production team sorts the possible questions into 15 bins one for each of the questions levels, using their judgment and experience to make them progressively more difficult
- being more educated doesn't help much