How we evaluate ourselves by comparing to known references

Leon Festinger and Merrill Carlsmith of Stanford University once asked their students to carry out an hour of excruciatingly boring tasks. They then divided the subjects into two groups. Each student in group A received a dollar (it was 1959) and instructions to wax lyrical about the work to another student waiting outside - in other words, to lie. The same was asked of the students in group B, with one difference: they were given $20 for the task. Later, the students had to divulge how they had really found the monotonous work. Interestingly, those who received only a dollar rated it as significantly more enjoyable and interesting. Why? One measly dollar was not enough for them to lie outright; instead they convinced themselves that the work was not that bad. Just as Aesop’s fox reinterpreted the situation, so did they. The students who received more didn’t have to justify anything. They had lied and netted $20 for it - a fair deal. They experienced no cognitive dissonance.

A bad excuse can be more persuasive than a good one.

Twenty dollars explained the lie. One dollar did not, so the students had to find another explanation. Apparently the work had been rather interesting after all. The mind balanced its books by quietly revaluing the asset.

We perform the same accounting with larger numbers. Buy an expensive product and favourable reviews suddenly look unusually rigorous. Work unpaid evenings for months and the project acquires historic importance. Defend a weak strategy in public and each disappointing result becomes proof that the breakthrough is now one quarter closer.

Behaviour produces evidence about identity: I chose this, therefore a sensible person must have had a reason. The more public or costly the choice, the more humiliating it becomes to discover that the reason was thin. Belief gets adjusted to protect the biography.

Changed opinions can, of course, be genuine learning. The suspicious moment is when conviction strengthens after commitment while the external evidence weakens.

Original criteria are useful witnesses. So are predictions written before the purchase, launch or grand declaration. Ask what an uninvolved person would conclude from the same facts, preferably someone who does not need yesterday’s decision to look heroic.

Cognitive dissonance can pull behaviour towards values. It can also employ belief as a full-time cleaner, polishing choices that should have been thrown out.