Book Cover


In Chapter 15 of Thinking, Fast and Slow we explore two types of fallacies. Logical and conjunction fallacies can impact any process improvement effort, typically in a manner that does not benefit change. 

The central plot device in this chapter is an experiment performed by Kahneman and Tversky that asked sets of respondents to rank attributes by representativeness and another group to rank by probability. The experiment begins with a description of the person, Linda,  (similar to the experiment at the center of Chapter 14). A set of statements about Linda’s potential profession is then listed. In this case (as compared to the experiment in Chapter 14), there are items in the list that require the application of logic to judge. For example, one item is “Linda is a bank teller” and a second is “Linda is a bank teller and is active in the feminist movement”. The lightbulb moment from the experiment was when there are two items that have a logical relationship, respondents distinguished between the two based on the story System 1 constructs. 

In the example, if we construct a Venn diagram in which one circle is bank tellers and the other is feminist activists,  the overlap would be small. Logic would dictate that there was a low probability of both being true (probability and plausibility do not mean the same thing). In the experiment, logic was overcome by representativeness (the story is perceived to be more plausible). In multiple runs of the experiment, almost everyone picked the entry “bank teller feminist” even though that was a very very very small proportion of the population. In the experiment, System 1 overrode System 2.

The failure of logic is a fallacy.  Kahneman defines a fallacy as a scenario when a person fails to apply a logical rule that is obviously relevant. He goes on further to define a  conjunction fallacy as a scenario where the conjunction of two events is judged to be more probable than one event if a direct comparison is made between events. Remember your Boolean logic!  Bank teller and feminist activist, both have to be true to be selected. 

The conjunction fallacy occurs when System 1 can construct a more plausible, coherent, or better story.  In the experiment, the story generates context so that System 1 creates a “better” story which overwhelms System 2.  Remember that we have established that System 2 is lazy and is less alert for inconsistencies when the story presented by System 1 is plausible. The conjunction fallacy can be avoided by becoming more specific.  Kahneman uses the comparison, “Mark has hair, Mark has blonde hair”. In this scenario, probability wins out with the majority of people selecting “Mark has hair” as being more probable. System 1 is overruled by System 2.

The experiment by Christopher Hesse comparing the valuation of two sets of dishes is illustrative of the impact of comparing different scenarios can have. IN the experiment, one set includes 24 pieces and the other set is a set of 40 pieces set but includes four relatively unimportant pieces are broken. When compared side-by-side the 40 piece set is judged to be more valuable, but when valued without comparison, people see the broken pieces and reduce the value. Most software groups are not making tableware but they often compare the outcome of different initiatives. For example, I overheard a team lead say that they had installed two pieces of work the previous evening. One piece of work was small and had no issues and the other one was very large and had a couple of minor issues. The fact there that were issue seemed to obscure the value of the large initiative.  All things being equal, logically the larger one was more valuable. In Hesse’s experiment removing the four broken pieces caused the value to go up (hence “less is more” in the chapter title). Clearly packaging and messaging are important to perceived value. With the conjunction fallacy, system one averages the two attributes on either side of the conjunction versus weighing them which is more of a system to function.

As I reread this chapter, I was reminded of how important owning the messaging is for change agents.  Knowing how to create comparisons that generate good conversations and highlight opportunities to improve is critical.  Change, an agile transformation, is significantly more difficult if the information shared prompts conversation based on logical fallacies.

Remember, if you do not have a favorite, dog-eared copy of Thinking, Fast and Slow, please buy a copy.  Using the links in this blog entry helps support the blog and its alter-ego, The Software Process and Measurement Cast. Buy a copy on Amazon,  It’s time to get reading!  


The installments:

Week 1: Logistics and Introduction

Week 2: The Characters Of The Story

Week 3: Attention and Effort

Week 4: The Lazy Controller

Week 5: The Associative Machine

Week 6: Cognitive Ease

Week 7: Norms, Surprises, and Causes

Week 8: A Machine for Jumping to Conclusions 

Week 9: How Judgement Happens and Answering An Easier Question 

Week 10:  Law of Small Numbers 

Week 11: Anchors 

Week 12: The Science of Availability 

Week 13: Availability, Emotion, and Risk 

Week 14: Tom W’s Speciality