Effect of the Survey on Company Performance

Moreover, whether a site s score improves or dechnes appears to depend on the actions taken at the site. In 1995 we inserted three questions to measure the site s response to the prior year s survey  

We computed a composite score for these questions and looked at the correlation between a site s response to the survey and the improvement in the survey score. There was a significant positive correlation (p . 03) between response to the 1994 survey, as measured by these questions, and improvement in the site s survey score from 1994 to 1997. This, of course, does not prove that the survey process was causative. It may be that effective safety-management systems are likely to respond to the survey and cause improvement in survey scores over time. 

About four months after beginning the implementation of the action plan, they saw a shift in the mean recordable rate from about 10 to about 4.2. This represents a 68 percent reduction. Another investigator, using our survey and a similar process, saw a reduction in the lost-time rate of over 50 percent (O Toole 2002). 

The information generated by analyzing the survey provides a clear basis for an interactive planning process. We have described the version we promoted and used for that process, but many other formats would serve quite weU. Our experience with the process defined here is that it is effective. We had success in changing the priorities of the company at the level of the executive committee, and the process was positively embraced by plant managers in the individual plant sites. 

---