Peter Klaas, Ph.D. Managing Director, WTT A/S

Donaldson’s (1987; 1999) Structural Adaptation to Regain FIT (SARFIT) model explains how structural adaptation is driven by the decrease in performance, or fitness, from misfits. The decrease in performance creates an incentive to remove misfits, so that performance can be restored and even increased from previous levels.

The SARFIT model as is operates on a single fit-misfit relationship between a structural and a contingency variable. In a multi-contingency context, for SARFIT to work, it is necessary to assume that misfits can be fixed independently, one at a time in a sequential manner.

With reference to complexity, a large body of theories and studies has argued that misfits may not be independent, in which case many or all misfits must be removed before fitness is regained. In this situation adaptation and change is revolutionary and rare rather than incremental and continuous. With complexity, SARFIT will work only by chance. In a test of their multi-contingency model, Burton et al. (2002) did not find that more misfits degraded performance more than just one misfit. With reference to complexity theory and Kaufmanns (1993) N,K model, they concluded that “rational changes in the design may or may not lead to improved performance, and may lead to large deterioration in performance” (Burton et al. 2002:1480).

However, for SARFIT to operate in a multi-contingency setting, complexity may represent less of a problem than what first meets the eye. Constraints reduce the number of misfits which must be considered simultaneously and offer opportunities for the designer to economize on information (Ashby 1963; March and Simon 1958; Alexander 1964; Arrow 1974).

In this presentation, we revisit the unique dataset on misfit from the Burton et al. (2002) study in order to learn more about the nature of misfits in general, and the potential impact from complexity and constraints on SARFIT in particular.