This is interesting, but probably not for the reason that the (apparently) telepathic authors think it is:https://eprints.qut.edu.au/64759/1/AUBEA_Yingbin_Feng.pdf
I myself am guilty of indulging in the odd bit of risk compensation. Deepest apologies.
"No, I will not work there until there is a guard rail".
Guard rail is erected.
"Yes, I'll work there now".
I suppose that is binary risk compensation.
Of course the question not answered in the study above is how many deaths or injuries there would be in the Australian construction industry if guard rails were never erected, nobody wore steel toe caps or face masks, or fall restraint systems hadn't been invented.
At least one poster here, apparently in all seriousness, believes that cycle helmets (and also hi-viz I think) should actually be banned. By law. It would be a fascinating experiment, and I would love to see the YouTube videos of the police enforcing the new "law". Still there would be a bit of "risk compensation" here, as I would stop cycling.
My personal experience as a manager in the construction industry was that those operatives more aware of risk and danger were also those who used PPE more frequently, and engaged in safer work practices. Generally older workers, having seen a thing or two.