According to Julie Nelson in her paper "Are Women Really More Risk-Averse than Men?", this question is like stating Canadians are right-handed, since ninety percent are. But, you would reject that thesis immediately. Studies demonstrate that women are more risk-averse than male counterparts, at least in the social laboratory; however, no study has shown that risk-aversion is feminine or quintessentially female.
The real question, when risk management and loss control programs are designed, is what risks are acceptable. Risk aversion may be the essence of great risk management.
The value of diverse opinions, risk tolerance, viewpoints, or just questioning standard operating procedures is enormous in retooling or creating a safety program in the workplace. Efficiency experts and quality control people often go to the first person along the assembly line and show them the finished product. They ask for input for improvements by way of the product, work conditions or safety. After carefully listening, they continue to move down the line getting more input. Why? Because these folks spend eight hours a day thinking about it. Nobody is more expert.
In the end, the management of the company has views from several stops along the assembly procedure, from divergent personalities, and from different levels of understanding the finished product. Inevitably, this source of information proves most valuable and accurate.
Now consider your safety program review. Everyone should be interviewed from the person ordering safety protection equipment to the president of the company. Don't ignore the office staff - very often are the best source for what really concerns the workers or how they view the company safety culture. Don't bias the study by interviewing only men or only women about certain jobs. You want to know the concerns of the most risk averse and the mentality of the least risk averse employee.
You may find the women to be very willing to take measured risks in some areas while men are very timid in the same. The important data involves what are the areas of agreement for what is universally acceptable risk and unacceptable risk. The middle is the range of risk aversion.
Regardless of the sex of the individual, some training may be appropriate to bring the most risk averse into the mainstream and the most risk tolerant into a safer mental environment.