We are not good at being objective

i have been working up some new material around the pattern basis of human intelligence. As a part of that I collected some of the common effects that are know to prevent objectivity in humans. At the moment I have these and thought I would share them:

• The contrast effect in which what we perceive is relative to what happened before. So if see a series of horrific images, then something you would in isolation see as horrific may seem ordinary or acceptable in sequence due to the contract.
• The sunk cost effect in which we are very reluctant to take a loss once we are committed, even if it is in our interested to do so, once committed we tend to stay committed
• Out group Homogeneity in which the in groups things they are diverse, but sees the outgroup as uniform. This leads to a desire for stereotypical models, we like guide books that say Arabs think like this when we would laugh at similar statements about our own society.
• Actor/Observer contrasts where we assume that we act in context, often accidently while THEY act deliberately. This is especially true if what we, or they do produces a bad outcome.   They intended evil, we just made a few silly mistakes.
• Self confirmation involves our ignoring any evidence that disturbs our pre-judgements or hypotheses and rationalisation means that we only search for data that will support the pre-judgements
• Confirmity means that we will alter our judgement to go with the flow, we observe the dominant players and attempt to support/follow them.
• Overconfident in our own abilities, we inflate estimates of our own capacity for judgement

Now those are known, and summarised in more detail in Thiele’s 2006 book The Heart of Judgement which is proving an enjoyable read. There are of course a lot more. What I find fascinating is that while the community as a whole knows of these errors they assume that the solution is to exhort or train people not to make them. In practice all of these are a consequence of the pattern basis and activation patterns through which we filter data. It follows from this that training people not to do the above will either not work or will only partially succeed which is more dangerous.

The other approach is to recognise that these are simply the downsides of what is otherwise a powerful evolutionary advantage of humans. If that is the case then we need to work with the effects above, not attempt idealistically to assume we can prevent them. The trainer who knows about the above, but then says they can train people not to exhibit those effects is themselves the victim of at least two!

More on this tomorrow.