Desirable change may appear chaotic, slow, or not sustainable. We may expect linear, continuous change, but it eludes us. Measurement and statistical analysis about behavior change often requires data showing continuity and having normal (i.e., Gaussian) distributions. When we encounter phenomenon that does not fit this expectation, we seek to transform the data to render it compatible with our method of analysis. We move from actual behavior to perception of it through surveys or transformations of the data. We believe this ignores “naturally” occurring data and what it says. Such techniques are convenient for statistics but may hide important features of the real phenomenon. Furthermore, desired behavior change is often nonlinear with a power curve distribution of the data. We explain why this occurs. We suggest how research and practice would be improved by using theories and methods that incorporate properties of non-normal distributions and discontinuous emergence.
The Journal of Applied Behavioral Science
Boyatzis, R. E., & Dhar, U. (2023). When Normal is Not Normal: A Theory of the Non-Linear and Discontinuous Process of Desired Change and its Managerial Implications. The Journal of Applied Behavioral Science, 00218863231153218.