ABSTRACT
The systemic-structural activity approach has developed principles of activity description as a multidimensional system and offered diverse models of task performance that capture various aspects of the activity structure (Bedny and Karwowski, 2003). This section describes a new method of reliability assessment of human performance that derives from the systemic-structural activity (SSAT) approach. Accuracy and reliability are two important characteristics of the system that are not the same. Accuracy refers to the precision with which a goal of the system is achieved, whereas reliability refers to failures of the system. An operator is a system component, and the accuracy and reliability of his or her performance influence the efficiency of the entire functioning system. There are a number of publications that cover a range of methods utilized for reliability analysis. However, there has been no attempt made to assess human performance reliability when a user interacts with a computer. The material presented here is the first attempt on using the task modeling method for this purpose. Moreover, a suggested method of reliability assessment can be used for reliability assessment of human performance in any system. In this chapter, we describe some basic concepts human reliability assessment of computer-based tasks from the SSAT perspective. Some methods that have been described in our previous chapters have been significantly modified and adapted to create a reliability assessment procedure that utilizes three stages including determining what kind of errors can occur and which errors can be considered as failures, what their probabilities are, and how these errors and/or failures can be reduced. This method not only permits to quantify human errors but also gives a qualitative description of such errors and suggests ways to reduce the number of errors and system failures. Existing cognitive psychology methods of assessing the precision of operator’s performance is not clearly separated from the methods of operator’s reliability assessment. These two methods of assessment are similar but
not identical. The basis for each approach is a method of analyzing human error but their purpose is different. Accuracy characterizes the precision with which the goal of task is achieved. Reliability refers to failures of performance and how the probability of failure can change over time or in stressful situations. Human performance can be precise but not reliable. Not all errors can be considered as failures. Some errors can be recoverable or have a relatively small effect on functioning of personal or technical components of the system. Other errors are associated with hazardous accidents, nonadmissible losses of time, and so on. Only the last category of errors can be categorized as failure. We will use the term errors to evaluate the precision of human performance and failures to evaluate human reliability. When accuracy declines and falls below acceptable level, it becomes an operator’s error. If as a result of operator’s errors the system cannot function and achieve its goal or goal achievement is conveyed by unacceptable losses, it is considered a failure. Therefore, the main criterion for distinguishing between errors and failures is their consequences for the system. There are errors or failures caused by technical components of the system and errors caused by operator’s erroneous actions. Of course, such division is relative. For example, an operator can perform wrong actions because the system design is not adequate or he or she does not possess the required skills.
