The Kevin Bacon theory
Kevin Bacon can be linked to anyone in the English-speaking world by six or fewer connections. If that is the case then anyone can be linked to anyone and they don’t have to be Kevin Bacon. The fraud prevalence rate for disability benefits are half a percent each, so we start with a 1 in 200 chance that each person knows someone committing disability benefit fraud but this doesn’t account for geographic demographics: someone committing fraud is probably poor and living in a poor area with other poor people. The rich are less likely to have regular contact with frauds, but even then the chances of actually knowing someone on the fiddle is very low.
But if the average person has 4 immediate family members- mother, father, brother, sister or more if they are a parent, then about 20 friends, neighbours and work colleagues, their chances of knowing at least one person committing fraud has drastically gone up even though there prevalence of 1/200 is still the same. It magnifies the appearance of the problem. Even if someone doesn’t know a person doing fraud, they are likely to know someone who does know a person falsely claiming.
So there are people who are really earnest and sincere in their anecdotes about people they know who are cheating the system. But they’re just not able to see how a small statistical prevalence can be compatible with many people knowing a fraud. This also doesn’t take into account just pure ignorance and maliciousness: the benefit fraud hotline gets about 600 calls a day and they’re nearly all rubbish.