> the argument is that it’s completely false

But is it completely false? Or does that 4% statistic also suffer from an over generalization based on very broad averages that can have extreme fluctuations when individual, specific cases are considered?