Take a look at the literature on "Risk Homeostasis Theory" (originated by Gerald Wilde). Some very fascinating work has been done in this area including some paradoxical results (e.g., improved roadway delineation reduces drivers' perceived risk when driving at night and leads to increased driving speed and, thus, can cause a net reduction in "safety").

Frank Schieber had a good comment that I second. "Perceived" safety often leads to more carelessness or ignoring regulations (often the things that actually provide more perceived safety). I've been a professional freight truck driver and a career Risk Management expert. Sometimes the best way to improve safety behaviors is to impress on the individual just how inherently dangerous a particular task IS. I'll take "tentative" and cautious over perceived safety and recklessness any day.

We did conduct before and after crash studies on a tidal flow lane project (where the centre lane changes direction between the morning and the afternoon periods) some 20 years ago. We designed the scheme to heighten a driver's sense of risk, so they would concentrate and obey the devices. We did find a reduction in crash frequency and severity when compared to the previous non-tidal flow arrangement. It sounds counter-intuitive to a point, but by increasing the driving task a motorist is more likely to drive to the speed limit, pay attention and not take additional risks. Such an approach can probably only be applied to relatively short sections of road and be effective. There is also some literature that has found when the driving task is too easy (boring), long-distance drivers are more prone to fatigue and fatigue- related crashes.

Thanks all for the helpful comments. Yes, I would concur with the comments that (over) confidence can lead to some increase in crash risk, and caution to increased safety. Frank, the risk homeostasis studies you highlight, and other research certainly suggest that we as drivers will drive to a certain level of risk (compensating for lowered risks by subconsciously increasing some other aspect of risk) or indeed novices driver training ...can lead to optimism bias and ...can unduly increase the confidence of novices about their driving abilities, leading to increased crash risk (See Lynam and Twisk 1995, and others in The effectiveness of driver training as a road safety measure - RACV report).

What about driver anxiety though rather than confidence per se? I'm thinking about a particular situation where drivers are not confident/anxious about using a particular route and prefer to use what they perceive as a safer route. Intuitively I think compelling them to use a route they feel anxious about may not be the safest option but I haven't seen any research on this.

I recommend checking the papers of Sam Charlton, such as this one: Article What’s the risk? A comparison of actual and perceived driving risk

I was not so confident as a driver although I used to have to drive for long distances. I think this made me cautious.

A nice 'driving memory' is that as I used to drive on small narrow country roads and always felt very honoured when tractors used to mount the low grassy bank at the side of the road when they looked at me and saw my navy blue nurse's uniform, and let me drive down the centre of the narrow one track road. This did not happen nearly so much when my husband and I were together in the same position but neither of us had a uniform on! They used to keep to one side and we would do the same, and pass each other very slowly and carefully.

Thanks all for the pertinent comments. I include below another article summary I received on the subject that might be of interest to the topic.

Driving fear and driving skills: Comparison between fearful and control samples using standardised on-road assessment

Joanne E Taylor, Frank P Deane, John V Podd

The present study explores driving skills in a group of 50 media-recruited driving-fearful and 50 control drivers, all of whom were women. Participants completed an on-road practical driving assessment with a professional driving instructor. Diagnostic as well as pre-post self-report and instructor driving assessments were conducted. Fearful drivers made more errors on the driving assessment than controls. However, the pattern of errors was identical for both groups, indicating that fear and anxiety may be associated with the number rather than the type of driving errors made. These differences remained when factors such as driving history, current driving frequency, and diagnosis were controlled using case selection. More research is needed to replicate the findings in more diverse samples. Additional work should also aim to clarify the specific role of driving skills in driving fear, which will facilitate treatment …

https://scholar.google.com.au/citations?view_op=view_citation&hl=en&user=OTwILP4AAAAJ&citation_for_view=OTwILP4AAAAJ:zYLM7Y9cAGgC