Iowa Gambling Task – Simulating real-life decision-making with card games
June 20, 2019
Decision-making is perhaps one of the most exciting capabilities of human cognition. At the same time, it also one of the most difficult ones to study as despite that our reasoning is often considered to be rational, the real-world decision-making is usually at least outwardly less so. To truly understand the cognitive processes behind decision-making, we need to be able to test and measure decision-making some way. One method for studying decision-making by simulating it is the so-called Iowa Gambling Task.
Originally named as just Gambling Task, the method was developed by a group of scientist from the University of Iowa Antoine Bechara, António Damásio, Hanna Damásio, and Steven Anderson. Some people might recognize the name António Damásio, whose Descartes’ Error: Emotion, Reason, and the Human Brain introduced the somatic marker hypothesis which has been extensively tested using this very method Damásio himself helped to develop. Somatic marker hypothesis will be discussed later as part of the criticism concerning the Iowa Gambling Task.
How Iowa Gambling Task works
The method itself is rather simple in how it works. Participants of the experiment are presented with four decks of cards and given an initial endowment of, e.g. 2000€. Participants are told that these digital decks all hold cards which will either reward them with an overall increase reward or penalize them through a decrease in it, and the name of the game is to try to win as much money as possible by choosing the right cards.
Now instead of using real cards, the cards in the experiment are often simulated on a computer screen. The reason for this is simple. The different decks differ from each other, how the rewards and penalties are balanced. Some of the decks have more bad cards and are therefore called “bad decks.” These bad decks, A and B, for example, have high rewards while also having higher losses, leading the expected value of the deck negative. Other decks are opposite of this, so-called “good decks” (e.g. C and D) with more tendency to reward the player and having a positive expected value by having lower losses but also lower rewards.
This difference between the decks is not told to the player, who will be left to form their own interpretation of the balance of the cards, which typical healthy subjects often succeed in doing.
Typical (and healthy) player often starts by sampling from the different decks, picking card here there. The player gets excited about the bad decks quite fast, but the high losses also quickly lead into the player mostly focusing on the C and D decks, of course occasionally tipping their toes back in the A & B decks but soon returning to the good decks. As the game ends, usually around 100th draw, the players are often unable to verbalize their tactic for this type of behavior, and the reason for this is simple. As the player progresses they will start to base their decision more implicit evaluation, often called gut feeling also, of which of the decks are good and which are not.
Measuring gut feeling
Once we pair the Iowa Gambling task with some form of physiological measurement of such as the galvanic skin response, how much your skin conducts electricity on any given moment, we can start to see the biological basis of the gut feeling. Measuring physiological changes also allows us to understand why some people are bad at the Iowa Gambling Task, and how emotions have a more significant role in our decision making that we’re accustomed to thinking.
Turns out, that while choosing a deck to draw a card from, the participants of exhibit an increase in electrodermal activity, meaning that they get physically excited and their sweating increases leading to increase the conductivity of their skin. Although this increase in sweating is minuscule, it is enough for measurement.
In the beginning, healthy participants will exhibit a skin conductance response (SCR) after drawing any large card, but eventually, the participants will develop anticipatory SCR that will occur soon as the card is about to be drawn and will grow stronger as the game progresses, especially when drawing from either one of the two bad decks. These skin conductivity responses become an index of gut feeling or intuition-driven decision making. The emotional excitement of drawing from risky decks starts to shape our decision process towards a more emotionally driven.
However, this anticipatory response is noticeable only if the participants are what we would call healthy. Participants with dysfunctions in the orbitofrontal cortex, brain area often considered to be highly involved in reasoning and decision-making, do not exhibit this type of anticipatory SCR at all. They never develop an intuition for choosing from the right card deck.
This finding is in line with what we often heard experts say about their areas of expertise. Chess masters, E-sports athletes, and even medical professionals often say that many of the decisions in their profession are made purely based on gut feeling. This is not that surprising as the search space for the right chess move, for example, is so large that without this type of emotion-driven learning of intuitive decision-making, developing any real skill in the chess would most likely be impossible. Or at least the games would be considerably longer, as making the correct decision would require large amounts of conscious processing and therefore also more time.
An argument against the Iowa Gambling Task
Most of the criticism towards the Iowa gambling task comes research focusing on Damasio’s earlier mentioned somatic marker hypothesis. The hypothesis proposes that emotional processes guide much of our behavior and decision-making. As we make decisions, for example, when choosing from which deck to draw a card from, emotions and physical changes such as excitement become associated with past choices and events. They develop into the so-called somatic markers. It is thought that these somatic markers are processed in the prefrontal cortex and amygdala, explaining why participants of the Iowa gambling task with dysfunctional orbitofrontal cortex (being part of the prefrontal cortex) are unable to develop an intuition for choosing from the right deck.
Scientists Barnaby D. Dunn, Tim Dalgleish, Andrew D. Lawrence in their paper The somatic marker hypothesis: A critical evaluation, criticize the somatic marker hypothesis and Iowa gambling task, arguing that the “anticipatory” response, for example, could actually reflect expectancies about the reward and penalty generated after a decision has been made. Iowa gambling task would, therefore, require alterations and variants of it, to control this methodological issue of causality. They argue that making the task less easy to comprehend consciously could be one of the ways of counteracting this.
However, despite their criticism, Dunn, Dalgleish, and Lawrence also point out that the Iowa gambling task has been incredibly generative paradigm as well, capable of being applied in a variety of experimental studies. Therefore there is no need to discard it as a research method, but instead, we should focus on improving it allowing us to understand how human decision-making even better and uncovering the real effect of emotions in our everyday actions and decisions.