The idea of 20mph limits in towns has already been shown to *increase* the accident rate presumably because pedestrians become less wary.
Citation needed.
Because that looks like seeing what you want to see in some study that might have involved special circumstances and other things being unequal.
The link posted a little later is to a source known to have many axes to grind and to spin shamelessly, and refers to one particular instance where a measure may (or may not) have been ill-considered in real life.