Everyday Chaos

This is a book review (Everyday chaos ), but you get the gist of David Weinberger’s main thesis – understanding causality can be very difficult – from this interview with Weinberger.

Weinberger suggests that it may not always be necessary or even most productive to wait until it is understood how things work. As the basis for this position, he points to the effectiveness of atheoretical A|B testing (let’s try this alternative and see what happens) and how AI can identify patterns that predict behavior and the mechanism responsible may not be obvious.

My initial reaction probably based in my training as a scientist was to label this as crazy talk. Much of science is about attempting to determine causal variables and manipulate or at least demonstrate how a causal variable leads to predictable consequences. However, as I thought about Weinberger’s position, I realized that there was some degree of commonality with my own thinking. When I taught the Introduction to Psychology course and talked about research in the field, I would sometimes make the claim that the scientists working in the physical sciences had it easy. This was partly to see if I would get a reaction. I argued that to understand the process of chemistry or physics, the researcher could manipulate something and did not have to worry about how the materials to which this treatment was applied interpreted this treatment or how they felt about reacting. Things are far more complex with human behavior in that the participant has a will and motives and history and these variables can moderate what happens. Using Weinberger’s concept, physical scientists have less chaos to contend with. With human behavior it is difficult to assure a given variable will act in isolation. 

My own research interests as an educational psychologist were more focused on applied research than the more basic science of many of my colleagues. I always thought that the basic scientists were kind of the “glamour boys/girls” of the field. They did the work that attempted to identify core scientific principles. We did the practical work of seeing if we could use these principles to generate outcomes that improved the human experience. I would argue that the transition from the lab to the classroom can be difficult. This may be reflected in the pattern of failed applications that advocates see as based on a logical basic principle. The repeated efforts to implement inquiry-based or learner centered tactics may represent examples. The philosophy proposing these ideas seems to gain traction every couple of decades and comparisons with direct instruction tend not to support these innovations. Maybe the issue is the inability to control how other variables interact with the core variables that are manipulated in applied settings. 

At some point, it may be possible to identify and control important variables that moderate the differences advocates focus on. However, without further insight and demonstrations based on these insights, implementations based on ideas rather than outcomes seem ill advised.