There is a lot of interest on the news channels for how the Scottish Government wants to frame the question for its proposed referendum on Scottish devolution.ï¿½ Does it really matter, as long as the question is clear?ï¿½ The answer turns out to be a might ï¿½Yes it doesï¿½ and this should result in a huge political tussle over the referendum wording.
Professional pollsters know this.ï¿½ When a client wants an opinion poll, the pollsters can frame their questions with great skill, to shift the results of the poll several or even many percentage points in one direction or another.ï¿½ For a marginal issue, a few points is all it takes to change the result.
Daniel Kahnemann is a towering figure in modern psychology (do watch him on TED if you have 20 minutes).ï¿½ He is a Nobel laureate in economics, because there is no medal for psychology, for research he did with late colleague Amos Tversky who sadly died before the Nobel Committee recognised the work.ï¿½ One fascinating experiment ï¿½ which I have reproduced myself in seminars ï¿½ demonstrates clearly the sensitivity of a result to the way we ask the question.ï¿½ In decision theory, this is sometimes called the ï¿½framing trapï¿½.
A terrible disease breaks out in a small village of 600 souls.ï¿½
It is expected to kill everyone and researchers identify two programmes that could fight the consequences:
ï¿½ will save 200 people
ï¿½ might save everyone ï¿½ with a probability of one third
ï¿½ or it might fail with no one saved ï¿½ with a probability of two thirds
Which would you go for?
Most people choose Programme A.
Now try two alternative options.ï¿½
Kahnemann and Tversky gave these to a different set of subjects to choose from.
ï¿½will result in 400 people dying
ï¿½ has a one third probability that no-one will die
ï¿½ and a two thirds probability that 600 will die
Now which one feels right?
Most people went for Programme B.ï¿½ This is odd, because if you compare the scenarios, Programmes A and B are identical in both examples.ï¿½ The difference is in how they are expressed.ï¿½ We tend to make choices based on a fundamental psychological predisposition to ï¿½loss aversionï¿½.
How you frame a question matters.ï¿½ When you ask: ï¿½Do you agree thatï¿½ï¿½ it biases people without strong views towards ï¿½yesï¿½ because saying ï¿½noï¿½ threatens the loss of something really valuable to us: affiliation ï¿½ fitting in.
Beware referenda.ï¿½ Always ask: ï¿½How would my answer differ if the question were framed in the opposite way?ï¿½