[belorn]

We should be careful when researcher themselves say that the data is unreliable. When it comes to false rape rates it is doubly so, as there is no common definition.

If we use the same definition as for conviction, ie proven beyond reasonable doubt that a crime did not occur, you get similar single digit. In one case it is the court that define proven beyond reasonable doubt and in the other the police, but they more or less is the same single digit of the total number of reported cases. We can assume that the police definition is a bit less strict than the court, but how much is just speculation.

We could use the rate in which prosecutor deem a case likely to succeed and thus brings it to the court and compare that to the rate which police finds it proven beyond reasonable doubt that the crime did not occur, but we will still end up with a huge rate of false reports.

Which is why people tend to compare the total number of reported cases vs those that police finds is proven (with evidence) beyond reasonable doubt that the crime did not occur. This makes the number of false report very low, and where the 2-10% comes from. To me that is just not very sincere approach.