Book Cover

Chapter 14 continues the discussion of cognitive biases and heuristics.  In Chapter 14 of Thinking, Fast and Slow we explore the representative heuristic.  

Tom W’s Specialty is an exercise/puzzle that Kahnamen uses throughout Chapter 14 to explore how the perception of representativeness impacts conclusions originally made from the base rate (the percentage of an attribute in the population). This might sound somewhat esoteric, but consider how many contracts are signed based on estimates that include representations of performance above and beyond the norm. Alternately how many agile teams pull more work items above their median performance based on representations of simplicity from stakeholders. The representative heuristic can negatively bias any intuitive decision. 

A base rate is the proportion of an attribute in a population.  For example, if a jar held 100 marbles and 40 where white and 60 where red, the base rate of white marbles is 40%. Knowing the base rate allows us to make decent predictions of the color the marbles if we drew 10 randomly. 

In the Tom W example, responders were asked to rank the probable program for a student would get a degree from.  Without other information, the responses matched the size of the student population in each program. When given a description of the student (even when they were told the description was not fully trustworthy) the pattern of responses changed.  System 1 thinking takes the additional data and fills in the blanks. 

Judgment based representativeness is often substituted for the knowledge that the base rate information and Bayesian statistics can provide.  The representative heuristic, like all heuristics, is a shortcut for making decisions based on comparison to a mental model. In the Tom W example, the experimenter provides a description of the person to prime model.  In the example of an agile team, the protestations of simplicity from the stakeholders trigger a mental model in order to modify what the team will commit to doing. Rather than using the base rate (the median number of work items completed per sprint in the past) the agile team thinks about the problem of how many things to commit to doing using a different model.  The use of the representative heuristic is a scenario where an easier question is substituted for a harder question. Instead of focusing on how many things can be done the focus is shifted to how easy the work will be. The representative heuristic injects bias.

Kahneman points out on page 152 that System 1 is not solely to blame when the representative heuristic takes us down the garden path.  He states System 1 makes the judgment while System 2 endorses it. Intuition is valuable only when constrained by the logic of probability. If you believe that there is a 40% chance of rain in the middle of the Gobi desert tomorrow, knowledge of probability suggests that you will be wrong.

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 Introductionhttp://bit.ly/2UL4D6h

Week 2: The Characters Of The Storyhttp://bit.ly/2PwItyX

Week 3: Attention and Efforthttp://bit.ly/2H45x5A

Week 4: The Lazy Controllerhttp://bit.ly/2LE3MQQ

Week 5: The Associative Machinehttp://bit.ly/2JQgp8I

Week 6: Cognitive Easehttp://bit.ly/2VTuqVu

Week 7: Norms, Surprises, and Causeshttp://bit.ly/2Molok2

Week 8: A Machine for Jumping to Conclusionshttp://bit.ly/2XOjOcx 

Week 9: How Judgement Happens and Answering An Easier Questionhttp://bit.ly/2XBPaX3 

Week 10:  Law of Small Numbershttp://bit.ly/2JcjxtI 

Week 11: Anchorshttp://bit.ly/30iMgUu 

Week 12: The Science of Availabilityhttp://bit.ly/30tW6TN 

Week 13: Availability, Emotion, and Riskhttp://bit.ly/2GmOkTT