This post will continue our thoughts re the use of safety culture surveys. The Oxford Dictionary says reality is the state of things as they actually exist, rather than as they may appear or may be thought to be. Another theory of reality is that there is no objective reality. Such belief is that there simply and literally is no reality beyond the perceptions, beliefs and attitudes we each have about reality. In other words, “perception is reality”. So, when a safety culture surveys is conducted, what reality is it measuring? Is the purpose of the survey to determine an “objective” reality based on what an informed and knowledgeable person would say? Or is the purpose simply to catalog the range of perceptions of reality held by those surveyed, whether accurate or not? Why does it matter?
In our August 11, 2010 post we noted that UK researcher Dr. Kathryn Mearns referred to safety culture surveys as “perception surveys”, since they focus on people’s perceptions of attitudes, values and behaviors. In a followup post on August 27, 2010 reporting some followup communications with Dr. Mearns we quoted her as follows:
“I see the survey results as a ‘temperature check’ but it needs a more detailed diagnosis to find out what really ails the safety culture.”
If one agrees that surveys are perception-based, it creates something of a dilemma as to which reality is of interest. If “things as they actually exist” is important, then surveys alone may be of limited value, even misleading, without thorough diagnostic assessments, which is Dr. Mearns' point. On the other hand, if perception itself is important, then surveys offer a window into that reality. We think both realities have their place.
We find some empirical support for these ideas from the results of a recent safety culture assessment at Nuclear Fuel Services.* The report is quite lengthy (over 300 pages) and exhaustive in its detail. The assessment was done as part of a commitment by the owners of Nuclear Fuel Services (NFS) to the NRC and in response to ongoing safety performance issues at its facilities. The assessment was performed by an independent team and included a safety culture survey. It is the survey results that we focus on.
In reporting the results of the survey, the team identified a number of cautions as to the interpretation of NFS workforce perceptions. The team found that survey numerical ratings were inflated due to the lack of an accurate frame of reference or adequate understanding of a particular cultural attribute. This conclusion was based on the findings of the overall assessment project. The team found the workforce perceptions to be “generally (and in some cases significantly) more positive than warranted” (p. 40) or justified by actual performance.
We found these results to be interesting in several respects. First there is the acknowledgment that surveys simply compile the perceptions of individuals in the organization. In the NFS case the assessment team concluded that the reported perceptions were inaccurate based on the team’s own detailed analysis of the organization.
Perhaps more interesting was that this inherent subjectivity of perceptions was attributed in this project to the lack of knowledge and frame of reference of the NFS staff, specifically related to standards of excellence associated with commercial nuclear sites. This resonates with an observation from our August 23 post that “workers who had been through an accident recognized a relatively safer (riskier) environment better than workers who had not.” In other words, people’s perceptions are influenced by the limits of their own experiences and context. Makes sense.
The NFS assessment team goes on to indicate that the results of a prior safety culture survey a year earlier also are compromised based on the very time frame in which it was administered. “It is reasonable to assume that the survey numerical ratings would have been lower if the survey had been administered after the workforce had become aware of the facts associated with the series of operational events that occurred” [prior to the survey]. (p. 41) We would add there are probably numerous other factors that could easily bias perceptions, e.g., people being sensitive to what the “right answer” is and responding on that basis; complacency; the effect of externalities such as a significant corporate initiative dependent on the performance of the nuclear business; normalization of deviation; job-related incentives, etc.
We think it is very likely that the assessment team was correct in discounting the NFS survey results. The question is, can any other survey results be relied on absent independent calibration by detailed organizational assessments? We will take this up in a forthcoming post.
In our August 11, 2010 post we noted that UK researcher Dr. Kathryn Mearns referred to safety culture surveys as “perception surveys”, since they focus on people’s perceptions of attitudes, values and behaviors. In a followup post on August 27, 2010 reporting some followup communications with Dr. Mearns we quoted her as follows:
“I see the survey results as a ‘temperature check’ but it needs a more detailed diagnosis to find out what really ails the safety culture.”
If one agrees that surveys are perception-based, it creates something of a dilemma as to which reality is of interest. If “things as they actually exist” is important, then surveys alone may be of limited value, even misleading, without thorough diagnostic assessments, which is Dr. Mearns' point. On the other hand, if perception itself is important, then surveys offer a window into that reality. We think both realities have their place.
We find some empirical support for these ideas from the results of a recent safety culture assessment at Nuclear Fuel Services.* The report is quite lengthy (over 300 pages) and exhaustive in its detail. The assessment was done as part of a commitment by the owners of Nuclear Fuel Services (NFS) to the NRC and in response to ongoing safety performance issues at its facilities. The assessment was performed by an independent team and included a safety culture survey. It is the survey results that we focus on.
In reporting the results of the survey, the team identified a number of cautions as to the interpretation of NFS workforce perceptions. The team found that survey numerical ratings were inflated due to the lack of an accurate frame of reference or adequate understanding of a particular cultural attribute. This conclusion was based on the findings of the overall assessment project. The team found the workforce perceptions to be “generally (and in some cases significantly) more positive than warranted” (p. 40) or justified by actual performance.
We found these results to be interesting in several respects. First there is the acknowledgment that surveys simply compile the perceptions of individuals in the organization. In the NFS case the assessment team concluded that the reported perceptions were inaccurate based on the team’s own detailed analysis of the organization.
Perhaps more interesting was that this inherent subjectivity of perceptions was attributed in this project to the lack of knowledge and frame of reference of the NFS staff, specifically related to standards of excellence associated with commercial nuclear sites. This resonates with an observation from our August 23 post that “workers who had been through an accident recognized a relatively safer (riskier) environment better than workers who had not.” In other words, people’s perceptions are influenced by the limits of their own experiences and context. Makes sense.
The NFS assessment team goes on to indicate that the results of a prior safety culture survey a year earlier also are compromised based on the very time frame in which it was administered. “It is reasonable to assume that the survey numerical ratings would have been lower if the survey had been administered after the workforce had become aware of the facts associated with the series of operational events that occurred” [prior to the survey]. (p. 41) We would add there are probably numerous other factors that could easily bias perceptions, e.g., people being sensitive to what the “right answer” is and responding on that basis; complacency; the effect of externalities such as a significant corporate initiative dependent on the performance of the nuclear business; normalization of deviation; job-related incentives, etc.
We think it is very likely that the assessment team was correct in discounting the NFS survey results. The question is, can any other survey results be relied on absent independent calibration by detailed organizational assessments? We will take this up in a forthcoming post.
No comments:
Post a Comment
Thanks for your comment. We read them all. The moderator will publish comments that are related to our content.