You are currently browsing the monthly archive for February 2009.

Decision Science News gives a tip on generating less biased estimates of your own future behavior:

When asked to make a forecast 1) generate an answer under ideal conditions, then 2) generate your forecast. Though you’d think the ‘ideal conditions’ would skew your forecast upwards due to anchoring, it does not. In fact, it causes you to generate more realistic forecasts of your own behavior.

Which is not too difficult. I believe that the logic behind this is that simply asking someone to “forecast X ” generates extremely similar results to asking someone to “forcast X under ideal conditions”. Asking someone (or yourself) to do both makes the fact that “non-ideal” conditions apply to forcast 2 very salient and thus more likely to take that fact into account. I suspect this technique is also useful for things like forcasting project completion  times.