Predictions Are Hard, Especially About the Future

Here it is, Tuesday evening, and I am finally settling down to write another blogfest – this, like many weeks, despite having specifically placed this high enough up on my to-do list that I genuinely expected to be starting in the early afternoon. The problem – one I am sure we are all familiar with – is not the writing, but all the other things I had planned to finish beforehand, which took far longer than I had originally estimated they would. Fortunately, such difficulties are illustrative of this week’s topic of discussion – the planning fallacy and how to counteract it.

The planning fallacy is a simple psychological phenomenon: human beings’ predictions about the time needed to complete a future task are usually significant underestimations. In some cases, wild underestimations: for example, when construction began on the Sydney Opera House in 1959, it was expected to be completed by 1963, but the site was not actually finished until 1973. Daniel Kahneman and his partner Amos Tversky were the first to describe this phenomenon, more than forty years ago, and Kahneman writes about it in his wonderful book, Thinking, Fast and Slow. He explains it as a kind of optimism bias, a tendency of people to adopt the rosiest scenarios as they imagine how a task will proceed.  Later scholars added other nuances to this explanation.  One reason for this apparent optimism bias, for example, might be the self-serving human tendency, when considering similar past situations, to take personal credit for all the things that went right (and thus assume they will go right again in the future), but to attribute errors and delays to outside forces that they presume will not occur again.¹  Nassem Taleb, in his book Antifragile, suggests it may not only be a psychological phenomenon, but also a consequence of a natural asymmetry: whenever circumstances or events cause a deviation from a well-laid-out plan, chances are far greater that the disruption will lead to delay than to expedition, so that the sum total of all deviations would always be expected to be postponement.

How many times have we seen the planning fallacy in action amongst our students? Just in the past month, I have met with returning students, vowing to perform better in this coming spring semester, who base this determination on unaccountably confident projections of all the steps they will complete to do so.  I have worked with February bar examinees, noses to the grindstone, who despite their genuine efforts are finding themselves slipping behind their intended schedules.  Not every student suffers from this bias, of course, and many of those who experience the bias don't actually suffer for it, either because they start with ambitious goals that leave plenty of leeway or because they find the extra time and energy to offset their underestimated projections.  Still, every year brings a significant crop of students who do not perform as well as they might have, because they seriously underestimate how long it will take them to complete an essay test question, compile a useful outline, learn the rules governing a specific legal topic, research, draft, and edit a significant writing assignment, or attend to the demands of student organizations.

Fortunately, the psychologists and scientists who have studied the planning fallacy have suggested a few strategies that can be used to counteract it, and these strategies are easily adoptable -- or correspond to techniques already used -- by academic support professionals.  In his book, Kahneman suggests the use of reference class forecasting -- that is, making predictions of the time needed to complete a task based not on a person's (or an entity's) internal sense of how long it should take them, but on observations of actual outcomes in prior similar situations.  In other words, if I were going to build an opera house, I might start off by assuming I could get it done in a few years, but if I considered how long it took to build the one in Sydney (and of course in other locations), I should understand that it is likely to take more than a decade.  Many of us do something at least adjacent to this with our students already -- providing them with estimates about how long they should expect to take to complete a case brief, for example, or to study for the MPRE -- but the idea of reference class forecasting suggests that it might be even more powerful to refer specifically to prior performances by other students.  Instead of saying, "You should devote at least 24 hours," it might be more effective to say, "Last year, every student who devoted 4 hours a day, every Saturday and Sunday, for three weeks, completed this successfully."

Another suggestion is the use of the segmentation effect. It has been observed that a person's estimate of the total time it will take to complete a task will be longer -- and thus likely more accurate -- if they are asked to segment the task (break the task down into a number of sub-tasks), to estimate the time it will take to complete each sub-task, and then to add all those times together to come up with the total time.²  However, there is a cognitive cost to being mindful and particular enough to break complex tasks down into numerous sub-tasks, and, without help, this kind of approach may be hard to learn and sustain.  Fortunately, this is just the kind of help we can give, especially to inexperienced students who may not be able to envision how a long-term task can be broken down, or even what all the steps involved might be.  By providing students with a framework of what to expect, and encouraging them to think realistically about what it will take to build each part of that framework, we can help them to stay on track, or at least in the general vicinity of the track, by using the segmentation effect.

Finally, another tool that has been suggested to combat the planning fallacy is the implementation intention, a term coined by Peter Gollwitzer for a particular model of thinking about future actions.  Encouraging people to think specifically about when, where, and how they will act towards their goal tends to make them more likely to move forward steadily, and in a timely way, towards them.  For example, people who received a telephone call in which someone asked them what time they planned to vote, from where they would be heading to the polling place, and what they would be doing just before they left to vote -- all questions designed to prompt them to think about when, where, and how they would vote -- were more likely to vote than those who did not receive the phone call.³  The mental IF-->THEN statement (as in, "If I am aiming to take a practice exam, then I should get a copy of an old exam from the library on Friday") is the implementation intention that moves people apace towards their goals.  This, too, is something that academic support professionals do, or can do.  By querying students about the specifics of how they expect to achieve their long-term goals, we can induce them to map out their plans in advance, changing vague ambitions about what they would like to achieve into articulable steps (the implementation intentions) that they can follow methodically to their desired ends within the time they have available.

It is a natural human tendency to overestimate what can be done in a given period of time. By helping our students account for this tendency, even if we cannot help them complete everything, we can at least help them get in a position where they've done enough to succeed. 

[Bill MacDonald]

¹Roger Buehler, Dale Griffin & Michael Ross (1995) It's About Time: Optimistic Predictions in Work and Love, European Review of Social Psychology, 6:1, 1-32, DOI: 10.1080/14792779343000112.

²Forsyth, D.K., Burt, C.D.B. Allocating time to future tasks: The effect of task segmentation on planning fallacy bias. Memory & Cognition 36, 791–798 (2008). https://doi.org/10.3758/MC.36.4.791

³Nickerson, D. W., & Rogers, T. (2010). Do You Have a Voting Plan?: Implementation Intentions, Voter Turnout, and Organic Plan Making. Psychological Science , 21 (2), 194-199.