Time Series Models of Human Factors Dynamics

Abstract

The past is one of the best predictors of the present and future behavior of individuals or organizations. Few would argue with this principle, yet it is largely ignored in the human factors literature. Most data analysis assumes independence of successive observations, and most theories deal with simultaneous values of several variables. This paper illustrates the dynamics of the two major subjects of human factors studies: organizations and individuals. Several representative human factors variables are shown to exhibit dynamism. Implications of a dynamic viewpoint are that static `heories may be incomplete, traditional data analysis methods are inappropriate for some types of human factors data, experimental designs for dynamic variables will have few subjects and many observations per subject, and dynamic analysis will be useful in human factors.