Is this not the problem your evidenced research fails to actually address. At some point, people aren't interested in how many lives a 10mph decrease will save. If increasing it means 35 more people might die (I've made the figures up but bear with me), then the percieved risk is virtually the same. And it works the other way around if you reduce it by 10mph, the percieved intrusion of rights (for want of a better phrase), rises exponentially.
At some point your research needs to point out how to change the consensus so it matters to people.
Having a look at the document I found this.
Changing a speed limit from 70mph to 60mph results in 94 fewer lives lost.
Changing the speed limit from 70mph to 80mph results in 18 more lives lost.
I assume these figures are modeled based on some collected data. One can, therefore, look at that and suggest that the 10mph difference from 60-70 is considerably more dangerous then the one from 70-80.
One could then manipulate that to suggest that it would be better to drive between 70 and 80mph as fewer lives are lost in that range then at the 60-70mph range!
Stupid really but that is what modeling and statistcal manupulation can give you.
I will read the whole document later, when I have the ability to concentrate on screen reading, and see how another group's opinion differs from this group's opinions.
