Errors in Confidence and Calibration

Broken clocks may be right twice a day, but even the most reliable clocks still need occasional calibration. The same concept can be applied to human judgement and decision-making. Far from normative standards of rational decision-making, people rely on a vast array of heuristics and mental shortcuts to efficiently, and often predictably, arrive at certain conclusions. Although from the adaptive toolbox approach these shortcuts are often assets when decision-makers are presented with cognitive or environmental constraints, heuristics are also accompanied by a host of biases and consequences when used improperly. Despite their bad reputation for contributing to irrationality in decision-making, heuristics aren’t solely to blame. Misconceptions of probability, ignorance of base rates, and perception of difficulty are integral facets in determining confidence: another guiding factor in decision-making. Confidence, whether warranted or not, can influence not only the choice of an alternative, but also the level of commitment to an alternative. Thus, when decision-makers fail to properly calibrate with respect to existing base rates and proportions, overconfidence and underconfidence emerge. Overconfidence and underconfidence, like heuristics, present themselves in a myriad of everyday situations, like parallel parking and choice of workout intensity, but can result in serious consequences.

Parking in downtown Auburn has never been an easy feat. During especially busy times, like Friday and Saturday night, my friends and I have been known to circle the block numerous times just waiting for a spot to open. One fateful day, my friends and I decided we simply had to have Taco Mama for dinner. After spending around 20 minutes hunting for a spot, we found one. There was only one problem: it was parallel and firmly sandwiched between two large trucks. My friends, desperate for dinner at that point, firmly insisted that I could do it. I, on the other hand, was not so confident. Aside from watching a 30-second animated tutorial on Instagram a few months prior, I had absolutely no coaching in how to pull off such a maneuver. Despite the fact that none of my friends had ever expressed any confidence in their own ability to parallel park, I was utterly convinced that my parallel parking skills were somehow worse than average. An egocentric train of thought guided me to answer the question “how good are you?” rather than “how good are you relative to others” (Moore & Cain, 2007). If I had paused for a moment to consider prior base rates (namely the fact that I’d never seen someone my age parallel park) and remember that parallel paring is not an element of the Alabama driver’s license test, I would have probably had a more appropriate level of confidence. Miraculously, I managed to park the car without a scratch, but if my judgement had been better calibrated, we probably could have eaten dinner much sooner by overcoming my aversion to tricky driving maneuvers.

About two months later, my friends witnessed another outcome of my uncalibrated confidence when we decided to take a run to relieve stress. Although I hadn’t kept up with my weekly workouts from high school, since I was no longer involved in team sports, I was overconfident that I could keep pace with my friends. About 2/3 of a mile into the run, however, I realized my grave error in judgement. Not much later, I lapsed into a minor asthmatic episode from pushing myself too hard without building up any endurance beforehand. Perhaps if I had tried to work out before attempting a challenging run, I would have received more feedback on my fitness level and had more accurate knowledge of my limits. It also would have helped to remind myself that my new running buddies were former soccer players who received far more aerobic conditioning than I had as a volleyball player. Eventually, after jogging for a few weeks to prepare, I was able to keep pace with my friends, but obtaining feedback on my fitness level and conducting a more thorough social comparison beforehand would have definitely saved me from embarrassing myself on the running trail.

While humans frequently display signs of over/underconfidence in everyday choices, severe consequences can be avoided or mitigated when calibration techniques are taken into account. Although conflation errors caused by egocentrism would suggest that engaging in a social comparison with a reference group is important, the choice of reference group matters. Perceptions of the reference group’s abilities and the perceived difficulty of the task at hand are also weighted heavily in determining a decision-makers level of confidence. While taking a more statistics-based approach to adjusting confidence levels by referring to base rates and probability would likely lead to better calibrated levels of confidence, decision-makers must inherently invest more time and effort into their decisions to do so. Similar to heuristics then, decision-makers are faced with an underlying tradeoff between accuracy and effort until the point where increased information no longer equates to heightened accuracy. The question then becomes: in an effort to better calibrate confidence and increase rationality in decision-making, how do we strike a balance between too much information and not enough?

References

Moore, D. A. & Cain, D. M. (2007). Overconfidence and underconfidence: When and why

people underestimate (and overestimate) the competition. Organizational Behavior and Human Decision Processes, 103(2), 197-213. https://doi.org/10.1016/j.obhdp.2006.09.002