It is time to recruit an assistant. Let’s assume you know a lot of potential candidates qualified for this job and that the candidates are split equally between men and women.
Today, you are supposed to meet five candidates. What is the most probable repartition of genders for these candidates, in their order of arrival?
Since there is a lot of potential candidates qualified for this job, with 50% of men and women, meeting a male or a female candidate are independent and random events.
Therefore, the three scenarii are equally likely.
However, the third scenario ”looks more” random and, intuitively, we would like this third event to be more likely than the other two scenarii.
This heuristic view results from the tendency of our brain to look for causality in totally random events.
If we meet only male or female candidates, we will make various hypotheses.
We will think that this is because the description of the job was gender-biased or because this kind of job is becoming more female or male.
In short, we will find a reason to explain a series of events whose only explanation is chance.
Thinking fast and slow, D.Kahneman
Epstein, R. J. (2013). Has discovery-based cancer research been a bust?. Clinical and Translational Oncology, 15(11), 865-870