Estimating the likelihood or impact of an event is difficult for the following reasons:
- Confirmation bias – once we believe something to be true we screen what we see and hear in a biased way to ensure our beliefs are proven correct.
- Availability bias – the more quickly we can think of something, the more familiar we judge that thing to be. For example, air travel is often seen as more dangerous than road travel
- Overconfidence bias – we think we know more about a topic than we actually do
- The anchoring rule – when we are uncertain about what the correct answer is we make a guess by grabbing onto the most recent number we have heard and making some minor adjustments to it
- The rule of typical things – the typical winter day is cold and cloudy, so how likely is a particular winter’s day to be cold and overcast? Very.
The good news is that people can become better estimators through calibration training and by using relatively straightforward techniques. One such approach, borrowed from engineering projects, is to put price ranges on the cost of materials. See the Tatton Hill Academy Case Study, for example.
The cost range for an item should be broad enough to give you a 90% chance of being correct. So when you consider the lowest number for the direct costs of recruiting permanent teaching staff, there should only be a 5% chance the real number will be below it. When you think of the highest number, again there should only be a 5% chance the actual number will be above it.