When judging the probability of anything we weight the salience of examples which comes to mind, how strongly they make us feel, and also how well they fit our model for what we expect. The latter, when applied to assess the probability that an object or event belongs in a category, is called the “representativeness heuristic”. It results in our neglect of important information like the base rate and can lead us to draw illogical conclusions.
The most famous example is that of Linda:

Linda was educated in a liberal college and heavily involved in social justice activism and women’s rights.
Which is more likely?
1. Linda is a banker.
2. Linda is a banker and a feminist.

Most people select the latter option, but this makes little sense because the first option also entirely includes the possibility of the second. Here the representativeness heuristic is leading us astray.