JDM Under Pressure¶
Evangeline Hong & Isabelle Levent
Spoiler Alert: The following describes a scene from Season 2 of The Good Place.
In the season 2 finale of The Good Place, a hilariously indecisive character, Chidi Anagonye, is presented with a final judgment test to determine whether he is doomed to eternal salvation or damnation: to pick between two hats in an empty room with a timer. Following eighty-two minutes of agony, Chidi selects the brown hat and exits the room. The Judge swiftly condemns him to lifelong torture for taking so long, upon which Chidi comically responds: “But did I at least choose the right one?”
In the context of judgment and decision-making (JDM), Chidi exemplifies exactly what not to do. But setting aside the comedic exaggerations, Chidi’s struggles capture a fundamental tension in judgement and decision making (JDM) broadly, but especially under time pressure: balancing effort and accuracy. The effort-accuracy model captures the inherent tradeoff between how good a choice is versus how many resources are devoted to making that choice. The optimality of a choice (“accuracy”) evaluates the decision ex-post based on the outcome and the rationality of the decision-making process (Payne et al. 1996). In other words: looking back, did I make the best decision I could? The resources spent on a choice (“effort”) refer to how much cognitive energy and time were allocated to reach the final decision. Since investing more effort to learn and process one’s options tends to lead to smarter and better decisions, the effort-accuracy model proposes that there is a compromise between effort and accuracy that the decision-maker must accept. This balancing act is particularly relevant when effort is either costlier or objectively capped by time pressure.
JDM processes depend on three dimensions: the total amount of information processed, the selectivity in information processing, and the pattern of processing (Bettman et al. 1998). Following the framework of the effort-accuracy model, the decision-maker could maximize accuracy by optimizing each of these dimensions (maximum information processed with no selectivity using a normatively perfect JDM algorithm), but it would come at a steep cost in the amount of effort. As a result, JDM under time pressure forces decision-makers to take adaptive approaches towards one (or more) of these dimensions.
There are three main adaptive responses: changing the speed of processing, changing the amount of information being processed, and changing the pattern of processing (Payne et al. 1996). First, individuals may accelerate their processing by simply spending less time processing each piece of information. This can be accomplished by increasing the amount of attention exerted to information processing. But because the limitations of cognitive effort are finite, this adaptation is mostly a first or auxiliary response. Second, individuals may additionally reduce the amount of information processed. This can take the form of reducing the number of sources being processed or by reducing the amount of information absorbed per source. Since both responses can be implemented without strategizing, they simply reduce effort at the direct cost of accuracy. As a result, they tend to be the most helpful when the time pressure is moderate. When time pressure is high, however, individuals may shift their decision-making pattern to strategically modify both the speed and selectivity of information being processed. Specifically, individuals may shift from a depth-first processing (where alternatives are evaluated completely and then compared) to breadth-first processing (where attribute(s) of alternatives are evaluated and compared). This adaptation seeks to more efficiently allocate the limited resources to process information that can really make a difference in the accuracy of the subsequent choice.
Time pressure can also be considered in the realm of risky decision making: traders make decisions in split seconds as new information comes in, negotiators try to reach agreements before a deadline, bidders place bets up until the last minute (Kocher et al., 2013). Research on this topic has produced mixed outcomes. Some argue that that time pressure unilaterally results in decreased risk taking, regardless of the type of choices presented. Others argue that the decision making response is more nuanced: when presented with situations that would result in gains, people are more likely to be risk-seeking under time pressure. When presented with situations that would result in losses, however, people are more likely to be loss averse, relying on heuristics and choosing safer options (Young et al., 2012).
Outside the laboratory, risky decisions are not generally presented as gambles. When assessing risk, people often look for a risk-diffusing operator (RDO), an action that would reduce the risk of a situation: taking a vaccine when going to a country with malaria, getting insurance, making duplicates of important documents. An experiment was done investigating whether time pressure increases the search for an RDO (Huber & Kunz, 2007). Participants were presented with a situation in which they ran a turtle conservation reserve, and had the choice of breeding the turles on the beach, where the water quality was subpar, or on a small island where the water quality was better, but there was the possibility of water mites infecting the turtle eggs. Under time pressure, participants focused on negative consequences of the risky choice and asked more questions about possible RDOs. The tendency was to find a way to mitigate risk, not to reject it outright. This provides better insight to the decision making process in the real world than black-and-white gambles and probabilities.
Adaptation: Different Decision Making Strategies Under Time Pressure¶
While people generally do a good job of adapting their JDM strategies based on the context of the time pressure, failures do arise and often are a result of one (or both) of the following shortcomings. People may simply lack the ability to execute their selected strategy due to computational or memory-related limitations. Alternatively, they may lack the knowledge to assess the decision-making context or be unaware of the strategies available to them (Payne et al., 1996).
Under time pressure, decision makers need to adapt their strategies. Normative strategies, developed in economics, describe what decision makers should do in ideal situations with infinite time to calculate and sort through information. In time constrained situations, decision makers employ different strategies, such as heuristics or breadth-first (as opposed to depth-first) searches, that minimize the amount of information processed. A common normative strategy is the weighted additive value strategy. The decision maker considers all attributes for each option, assigns a weight and value to each attribute, and calculates the expected utility by summing the product of the attribute value and weight for each option (Payne et al., 1996).
Other common strategies that people use to guide their JDM process under time pressure include (Bettman et al. 1998):
table, th, td { border: 1px solid black; }
Name |
Explanation |
Benefits |
Lexicographic (LEX) |
1. Select the attribute you value the most 2. Evaluate this attribute for each of the options 3. Compare and select the option that scored the highest in this attribute |
Individuals are limiting the amount of information being considered, which reduces the effort associated with processing and computing |
Satisficing (SAT) |
1. Predetermine a cutoff standard for each attribute 2. Consider each alternative one-by-one 3. Evaluate whether every attribute of the alternative exceeds the cutoff threshold 4. Select the first alternative that meets this criterion 5. Reduce your cutoff standards and repeat if none of the options make the original cut |
Individuals can reduce the computational effort by simply making repetitive, simple comparisons while ensuring that the selected choice is not unacceptable in any way. |
Elimination-by-aspects (EBA) |
1. Select a few attributes that are most important to you 2. Predetermine a cutoff standard for each of these selected attributes 3. Consider each alternative one-by-one, evaluate whether every attribute of the alternative exceeds the cutoff 4. Select the first alternative that meets this criterion |
Combination of the benefits of LEX and SAT |
Equal weight additive (EQW) |
1. Evaluate each attribute of an option 2. Equally weight and sum these to get the subjective value of the option 3. Compare and select the highest rated option |
Closest to the normative strategy in comprehensiveness while reducing the computational effort in assigning different weights. |
Frequency of good/bad features (FRQ) |
1. Predetermine a cutoff standard for each attribute 2. Consider each alternative one-by-one 3. Evaluate whether each attribute meets the cutoff threshold 4. Count the number of attributes that do meet the standard 5. Select the option with the highest counts |
Combination of the benefits of SAT and EQW |
Consumer choice¶
Consumer choice refers to the selection of products and services by consumers and is a classic context in which decision-making has been studied. Behind each consumer decision, there is some degree of uncertainty (the utility to be derived from the purchase) and some need for timely decisions (either externally imposed in bargain settings or internally imposed as opportunity cost).
Consumer decisions under time pressure exhibit two main changes. The first and most prominent change in the JDM process was a shift from depth-focused to breadth-focused strategies (Payne et al. 1996). In other words, time pressure motivated consumers to cover as many alternatives as they could by focusing on a select number of attributes, rather than attempting to comprehensively evaluate each option before moving onto the next. This can be seen in eye-tracking studies that observed how consumers shopped under time pressure: shoppers began by quickly scanning each of the options for their respective brand displays (Pieters et al. 1997; Reutskaja et al. 2011). This strategic shift was not only evidenced in how consumers collected, but also in how they processed the information. When presented with a JDM task under time pressure, consumers felt much more confident about their decisions when they were presented with a breadth of information rather than a depth of information about a limited number of options (Jacoby et al. 1994).
Another notable change is increased selectivity in information processing. Under time pressure, consumers strategically prioritize particular attributes to evaluate rather than going through an exhaustive laundry list of product features. One common filtration mechanism is identifying attributes that can serve as a proxy representation of the other attributes. The most frequently cited examples of these representative attributes are brand name and price (Nowlis 1995; Suri and Monroe 2003). Consumers may assume that this feature reasonably reflects information about the quality, popularity, or other specific attributes that they no longer have to individually reason about.
Time pressure elicits many adaptations from individuals as they reason about their consumer choices: from computational shifts in approaching the problem (breadth over depth) to implementational shifts in executing the strategic shift (filtering of information). These adaptive JDM strategies seem to manifest intuitively, sometimes even without the decision-maker’s realization. For the most part, the use of these heuristics (such as the LEX or EBA rule) have been proven to increase the accuracy of time-constrained consumer choices relative to making the decision using a normative strategy. In some cases, however, JDM under time pressure breaks down. Perhaps the most extreme example of this is seen in the occasional preference reversal between choices made under time pressure and choices made in the absence of time constraints. While the stakes of a suboptimal consumer decision may not be so dire, JDM under time pressure can be found in contexts with higher stakes and under tighter constraints. To better understand how individuals take on these challenges, we turn to experts who operate and specialize in JDM under time pressure.
The Recognition Primed Decision Model¶
Gary Klein, a leading figure in the field of naturalistic decision making, proposed that experts use different strategies than novices and that decision making in the real world is different from decision making in a laboratory. He interviewed nurses, pilots, software engineers, firefighters, economics professors, and soldiers –- all experts in their respective fields -– and developed a new theory for expert decision making under time pressure: the Recognition-Primed Decision model. This model focuses on the experience of the decision maker, tactics such as intuition and mental simulation, and a decision-making strategy that emphasizes coming to conclusions without direct comparisons.
Klein emphasizes the importance of the decision maker’s level of expertise. Experts in a field are able to “see the invisible” (Klein, 2017, p. 151). Their experiences allow them to better pattern match, consciously or unconsciously, often resulting in gut feeling about the right decision. Nurses in a NICU were often able to tell when a baby was at risk of becoming septic and when to put them on antibiotics. When asked to explain her decision, one nurse listed out a series of symptoms and cues she looked out for, some typical textbook symptoms and others based specifically on her experience. Klein points out that the best decision makers also know that their intuition can be fallible, but years of experience allow them to notice positive and negative cues, categorize which events as typical or atypical, have broad situational awareness, and discern minute differences in given information or situations.
The Recognition-Primed Decision model relies heavily on mental simulations. Decision makers create mental models, usually containing no more than three key factors, to imagine what might occur in a given situation. Klein interviewed a lieutenant firefighter about his decision making process during the rescue of a woman who jumped off a highway overpass and was lying semi-conscious on a metal fence below. First, the lieutenant considered using a rescue harness. He imagined snapping it onto the woman and realized she might fall off the fence as they attempted to attach it. He discarded the idea, and then imagined attaching the rescue harness from the back. Again, the idea was discarded when he realized she could still fall off. He imagined another scenario, and then came up with the strategy of using a ladder belt, a strong belt that firefighters use, to hoist the woman. This processing all happens in a minute. He moved forward with the last plan, but as they lifted the woman, he realized that ladder belts are meant for bulkier firefighters. The woman fell, and was caught by the rescue team standing with a ladder below her. Mental simulations are not always accurate, but they allow an expert decision maker to quickly go through different scenarios, discarding them as needed. When a potential flaw is revealed through the mental simulation (ex: a fault in attaching the rescue harness), the decision maker moves onto imagining the next scenario. This follows the satisficing approach of decision making with the decision maker selecting the first option that works. Klein argues that for expert decision makers, the first decision is often the right one (Klein, 2017).
Conclusion¶
People are under time pressure all the time in their daily lives both personally and professionally. The studies discussed indicate that time pressure impacts decision making, whether or not one is always cognizant of the change. Normative strategies that are supposed to result in the best decisions perform poorly under time pressure. In fact, in Payne’s classic experiment of JDM under time pressure, the normative strategy underperformed every other possible strategy when participants were choosing between different gambles (Payne et al., 1996). Time pressure can also result in preference reversals as seen in an experiment with college students choosing housing; the students preferred one type of housing over the other under time pressure, and the opposite without the time constraint (Svenson & Edland, 1987). Strategies that increase performance under time pressure include EBA and LEX both of which emphasize selecting one or a select few attributes to focus on and conducting a breadth-first search (Bettman et al., 1998). Using these heuristics, one is able to get through more options in less time, even if each option is evaluated less rigorously. In contrast, experts, possibly some of the best decision makers, rely not on simple heuristics, but on years of experience. Under time pressure, their knowledge gives them an edge. They are able to create mental simulations, rely on instinct, and come to conclusions about possible choices more quickly (Klein, 2017).
References¶
Jacoby, J., Jaccard, J., Currim, I., Kuss, A., Ansari, A., & Troutman, T. (1994). Tracing the Impact of Item-by-Item Information Accessing on Uncertainty Reduction. Journal of Consumer Research, 21(2), 291-303. http://www.jstor.org/stable/2489821
Kocher, M. G., Pahlke, J., & Trautmann, S. T. (2013). Tempus Fugit: Time Pressure in Risky Decisions. Management Science, 59(10), 2380–2391. https://doi.org/10.1287/mnsc.2013.1711
Nowlis, S.M. (1995). The effect of time pressure on the choice between brands that differ in quality, price, and product features. Marketing Letters 6, 287–295. https://doi.org/10.1007/BF00996192
Payne, J. W., Bettman, J. R., & Luce, M. F. (1996). When Time Is Money: Decision Behavior under Opportunity-Cost Time Pressure. Organizational Behavior and Human Decision Processes, 66(2), 131–152. https://doi.org/10.1006/obhd.1996.0044
Reutskaja, E., Nagel, R., Camerer, C., & Rangel, A. (2011). Search Dynamics in Consumer Choice under Time Pressure: An Eye- Tracking Study. The American Economic Review, 101(2), 900-926. http://www.jstor.org/stable/29783694
Suri, Rajneesh & Monroe, Kent B. (2003). The Effects of Time Constraints on Consumers’ Judgments of Prices and Products. Journal of Consumer Research, 30(1), 92-104. https://doi.org/10.1086/374696