Mechanisms of human error and human error detection.

摘要

This thesis explores the underlying mechanisms of the execution and evaluation of human action by examining human errors and how they are detected. The thesis is divided into two parts. Part I is a theoretical discussion consisting of three chapters. Chapter 1 provides background and a literature review on human error. Chapter 2 is a discussion of problems in methodology for error research. The third chapter is concerned with theoretical issues surrounding the detection of human error and includes a review of the relevant literature.;Part II describes four experimental studies of error and error detection. The first three studies focus on error detection. Chapter 4 is a diary study of everyday errors with emphasis on the detection and correction of these errors. From this comes a general taxonomy of detection mechanisms. Chapters 5 and 6 examine detection of errors in the context of a keypressing task and skilled typing, respectively. These two studies address detection mechanisms on the basis of response-produced feedback only. In both experiments, the tasks were designed such that the results of actions were not available for examination by the people who committed them. Finally, Chapter 7 reports an experiment which illustrates how the design of the interface between human and computer can affect the frequency of particular kinds of errors.