top of page

Breaking with time and space

How to make better decisions under uncertainty (Part 3)

By Pedro Del Carpio.


In the first article of the series, we discussed that improving our decision-making process using the Subjective Expected Utility (SEU) method requires us to identify the courses of action we can take, the outcomes and their probabilities, and estimate the level of (dis)satisfaction of their consequences (the utility). Theoretically, once these elements have been accurately established, the SEU will result in a straightforward answer (Figure 1).



That being said, if you have already tried to put this method to use, you probably found that it’s far from an effortless task. Although the arithmetic knowledge needed for the required computation is rather basic, estimating trustworthy values of the elements mentioned above is in many cases quite a challenging endeavor.


As a way to facilitate this process, so far we have suggested using external data to adjust our forecasts using the external view and updating our probability estimates via Bayes' Theorem. If correctly applied, they can reduce to some significant degree the level of uncertainty of the case we have at hand.


Continuing with this goal, this chapter of the How to Make Better Decisions under Uncertainty series will expand on two powerful decision-making aids. First, we’ll explore the way in which thinking that something has already happened and failed can help us unveiling so far hidden alternative courses of action. Secondly, we’ll show how using other people’s lives lets us gain better insights about the future desirability of our actions’ consequences.


Premortem: Explaining the future past


You are probably familiar with the concept of a postmortem, the examination of a corpse to determine the cause of death.

This concept is also used in fields that deal with risk management, such as the military or in the business world, where a postmortem is conducted to explain why a past event has succeeded or failed. This procedure has proven useful when trying to learn from the past and to improve future decisions.


Research psychologist Gary Klein suggests going one step further and apply what he calls a premortem method [1]. A premortem takes us to the future — shifting the direction of the previous logic — exploiting our cognitive capabilities to makes us imagine that the event has already occurred and explain why it failed. Unlike a postmortem, which is performed after the resulting outcome of a decision has happened, the premortem is done before any commitment is made, allowing us to decrease the likelihood of a forecasting blind spot that could lead us to a costly action.



decision making optimization, prospectiva retrospectiva, behavioral sciences tool
Photo by Andalucía Andaluía on Unsplash

The act of explaining future events as if they had already happened is called prospective hindsight (Figure 2) and it’s facilitated by our natural tendency to come up with causal chains for past scenarios. This point is key. Studies have shown that with the benefit of knowing the outcome, subjects are prone to find initially unpredictable events as predictable and inevitable [2]. Under this mental state, humans are very confident about the explanations we make after the fact, falling into an illusion of understanding the past known as the “I knew it all along” effect. Interestingly, with the premortem method, this phenomenon called hindsight bias  [3] can become a useful tool.


When our goal is knowing what the possible outcomes of an action could be, imagining that it already happened and that it went wrong, forces us to see its pitfalls in hindsight; and therefore, they become more evident than when we try to imagine what could happen in a situation that hasn’t happened yet. As part of their pioneer research on the topic, Mitchell et al. [4] concluded that “a backward perspective on the outcome to convey its certainty should generate more insight into events that might, in fact lead to it” (p.36).



Thus, this exercise allows us to see with more clarity previously ignored potential consequences and decreases the chances of very likely overconfident estimates. For example, suppose you are back in the situation where you have to decide what path to follow after finishing high school. The standard set of possible courses of action is:

a) Choose a career I am passionate about.

b) Choose a career that is less interesting but better paid.


However, we can still discover reasonable additional courses of actions by using the premortem method. To do so, you have to imagine you already completed your studies, started your professional career and things went terribly wrong. Then determine the causes that made this choice a failure.


Some of your answers could look like the following:


“I wasn’t really sure about what I wanted.”


or


"I didn’t put in the required effort because I didn’t like working in that field.”


This prospective hindsight exercise lets us produce additional courses of action. For instance, a new alternative that could be contemplated when making a decision with the SEU method is “Delay my studies and get some work experience to explore what I like.” Considering this additional possibility might decrease the probability that the case in hand fails.


Therefore, mentally time traveling to identify alternative courses of action lets you see beyond the evident options (frequently incomplete and prone to biases), modify the direction of your plans if needed, and moderate any overly optimistic estimates.


Substituting: Put yourself in the other's shoes


To make decisions we must estimate how we will feel — how satisfied or dissatisfied about the consequences of events that haven’t happened yet. In economics terms, we need to forecast the utility we will get from our actions. This process comes naturally to us: we use our imagination to create a simulation of how the future will play out, and we assume that the emotions we foresee are an excellent approximation of what is really going to happen.


Although this estimation is on many occasions sufficient enough for us to make routine and mundane decisions, in many situations it can lead to significant errors. In a vast number of circumstances we misjudge how we will feel when we get what we think we desire. Quickly recall, how many times have you made a decision, just to end up feeling surprisingly disappointed when you got what you want?


This cognitive shortcoming resides in the fact that our minds inevitably construct an image of a misshapen future loaded with add-ons and subtractions. In the best-selling book Stumbling on Happiness, psychologist Daniel Gilbert suggests the following reasons [5]:

  • We make mistakes remembering our emotions associated with past events, therefore failing to accurately recall how we would feel if these situations happened again.

  • We imagine a future that can only happen as we foresee it, not realizing we are overlooking other relevant possibilities.

  • Our current emotional state profoundly influences how we think and feel about the future. When the actual circumstances change, we are prone to also change our forecasts.

  • Our foresight abilities ignore that things will look and feel different once they happen.Our foresight abilities ignore that things will look and feel different once they happen.


As a mean to overcome this limitation, instead of merely relying on our intuitive prediction of future emotions when we want to decide what path is more appealing, Gilbert proposes using the Surrogating approach, which implies using the experiences of people that happen to have lived or are living the scenario we are considering.


Humans have the luxury of being able to acquire data by capitalizing on others’ knowledge and emotions, avoiding the necessity of spending all our resources on trial-and-error learning. In order to know if you would enjoy something, you don’t have to actually change the car brand you drive, move to that new trendy neighborhood in town, or start living as a digital nomad yourself. We can use what people who are living that experience have to say, giving us access to information more accurate and complete than what our imagination is capable of coming up with.


sustitución, behavioral sciences tool
Photo by Rob Curran on Unsplash

The basic execution of this approach is well known. When facing these kinds of dilemmas we first tend to ask our family, friends or acquaintances about their opinions, which is a fair elementary solution. However, the big limitation with this tactic is that as individuals, they also suffer from cognitive biases and memory recollection limitations. Additionally, it’s not uncommon to receive information shaped to confirm our prior beliefs and makes us feel good, whether it is warranted or not.


Therefore, whenever possible, search for the input of competent third-party sources, such as experts and people who have actually lived the experience and have no personal attachments to you. The advantage of vicariously learning is not anecdotal. Research has shown that the estimation of future happiness based on the report of a randomly selected person who already lived the experience (surrogating) is more accurate than relying on our predictions (simulating)[6].


Luckily, the Internet has dramatically facilitated the surrogate process, allowing us to directly learn from subjects who have already tried what we think we want. With the click of a few buttons, review websites including Tripadvisor, Yelp, or Amazon show us the opinions and scores given by thousands of customers eager to share their experiences with hotels, restaurants, venues or products. Estimates suggest that 77% of people read online reviews before making a purchase, and that 9 out of 10 people think that reviews are as important as personal recommendations. Here, supported by the law of large numbers [7] we can be confident that the more reviews a product or place has, the closer we will be to a genuine glance of what our opinions and feelings will be if we get to have the same experience.


Another excellent source for surrogating is Quora, the question-and-answer site where all kinds of topics are asked, answered and organized by its members. Want to know if getting rich is worth it? Is retiring very early (30’s) fulfilling? Or what it is like to quit your job and travel the world? Give it a try. Although exposing ourselves to this information can’t tell us what to do, it will certainly expand the range of our emotional repertoire by accessing the inner world of people “who have been there”, with the byproduct of improving our decision-making process.


Main takeaway


This article explained two ways in which we can reduce the limitations we face when using the Subjective Expected Utility method for decision-making. The premortem analysis allows us to generate alternative courses of action via the prospective hindsight exercise. With the Surrogating approach, we use the experience of others to improve the estimate of our feelings about the consequences of a future event.


Notes


[i] Because there is always the possibility that the states of the world will change due to the effects of time, regardless of whether we employ USE correctly, we can make decisions that lead to undesirable consequences. Even with USE it is impossible to control the vagaries of fate.


References



[2] Hindsight Bias from 3 to 95 Years of Age by Bernstein, Erdfelderm, Meltzoff, Peria y Loftus


[3] Hindsight Bias by Roese and Vohs



[5] Stumbling on Happiness by Daniel Gilbert


[6] Surrogation: An Antidote for Errors in Affective Forecasting por R. J. Norwick, D. T. Gilbert, y T. D. Wilson (unpublished manuscript, Harvard University, 2005).


[7] The law of large numbers via Encyclopedia Britannica





Comments


bottom of page