Metacognition in Dating:
Coffee Meets Bagel Case Study
Foundations in Human Factors, April 2017
Introduction
“The wise man knows he knows nothing, the fool thinks he knows all” (African proverb). The acknowledgment and judgment of one’s own knowledge is not only philosophical, it is indicative of human metacognition- the ability to think about thinking. Metacognition is a process that runs parallel to cognition, enabling individuals to set goals, appraise their own abilities, develop strategies, and monitor and adjust strategies in order to maximize learning and decision making (Nelson, 1990; Shimamura, 2008). Research suggests that metacognition has a wide-reaching role, influencing reading comprehension, writing skills, memory, problem solving, self control, self instruction, and many other areas (Flavell, 1979).
Unlike computers, humans are not perfectly rational. Humans often do not process data effectively to generate the optimal solutions to their problems. Challenged with limited cognitive resources to address complex problems, humans use heuristics that maximize efficiency, but lead to biases and errors in judgment. To counteract these challenges, designers must leverage metacognition by enhancing the user’s ability to self-direct and orient themselves towards successful interaction decisions. This design review will examine the metacognitive factors present in decision-making in the dating app, Coffee Meets Bagel. Design recommendations will be provided to enhance user ability to self-appraise, develop strategies, and modify these strategic decisions in the pursuit of goals and expectations.
Metacognition Literature Review
Metacognition is an essential part of the human ability to adapt top down, cognitive abilities with information from the environment for efficient and effective output. Goals and self-evaluation are the starting blocks of metacognition. Individuals set goals in order to learn and develop competence (Paulson & Bauer, 2011). Clarity, challenge, commitment, feedback, and task complexity all contribute to a person’s motivation to reach a goal (Locke & Latham, 2006). After goals are set, people engage in introspection to determine their position in relation to their goals. They identify strengths they may leverage, weaknesses that must be improved, and the relative intensity of these strengths and weaknesses on the continuum of good to bad. They may assess their learning style, memory capabilities, and shortcomings to target areas for improvement (Paulson & Bauer, 2011).
Strategy is how an individual’s goals and self-appraisal are translated to action. However, different situations may have high degrees of complexity and previous strategies may be ineffective when applied to new contexts (Waters, 2010).
For strategic changes to occur, strategies must be monitored through metacognition. Nelson and Naren’s (1990) model of metacognition proposes that meta level processors receive and evaluate input from cognitive, object level processors. Feedback control is initiated based on this evaluation. Individuals monitor their progress towards goals, errors made, the overall effectiveness of the strategies in use, and then revise or change strategies accordingly. Often, this monitoring is regulated by metacognitive experiences that include judgments of learning, feelings of knowing, and confidence ratings of previous performance (Flavell, 1979; Shimamura, 2008).
Subtopic: Decision Making
In decision-making, we constantly use metacognition to assess goals, evaluate complex information, activate prior knowledge, and modify goals based on progress. These decisions have wide ranging implications on what, where, how, and with whom we live our lives. They influence our relationships, quality of life, and prospects for the future.
Early models of decision making assumed that decision makers appropriately understand and weigh all decision options and possible outcomes to compute maximum utility and value (Edwards, 1954). Utility theory was proposed as a modification to this model that accommodates individual preference. It contends that decisions are a product of an individual’s subjective ratings of utility of the individual decision choices, combined with the probability (Fishburn, 1970).
However, human decision-making in reality is less formulaic than the early models of decision making imply (Tversky & Kahneman, 1981; Sternberg, 2010). According to Simon’s concept of bounded rationality, decision making is limited by several factors- the intrinsic nature of the decision problem and goal, the availability of properties of the external world, e.g. time and resources, and the limitations of the individual mind (Simon, 2000).
But how do these factors interact? Kahneman and Tversky’s (1979) seminal work in behavioral economics suggests that people’s responses to risk and uncertainty do not follow the normative models of rational choice predicted by utility theory. Research demonstrates that people are irrational, favoring certain, smaller gains over uncertain, but probable, larger gains (Kahneman & Tversky, 1979). People also tend to disregard features that are shared by all choices under consideration, which leads to inconsistencies when the same choice is presented in an alternative context. As a result of these findings, Kahneman and Tversky developed Prospect Theory, which suggests that decisions are made as a function of the value of gains and losses. These decisions are weighted lower than corresponding probabilities would be, with the exception of low probabilities, which are overweighted. This exception amplifies the value of long shots and increases the adverseness of unlikely, severe losses (Kahneman & Tversky, 1983).
Decision making theory indicates that the mind must seek strategic ways to rationalize problems and seek the best solution among the possibilities given the provided constraints. However, this rationalization is bounded by an individual’s subjective level of ability and the inherent limits to human information processing, which lead to imperfect mental shortcuts.
Subjective Factors in Decision Making
Real life decisions do not occur in a vacuum. External and internal factors may have a considerable influence on people’s strategic decisions (Waters, 2010). Metacognition must be invoked to recognize and facilitate a rich environment for the achievement of goals.
Individuals vary in their cognitive abilities and learning styles, influencing their levels of working memory and attention, ability to manage knowledge, deal with consequences, strategize, and manage their wants and needs (Simon, 2000). As a result, the elderly and people with cognitive disabilities are also often particularly challenged during decision-making. Expertise also influences the structure of metacognitive decisions. Novices must expend more time and effort than experts to monitor their strategies and progress towards goals, as they must simultaneously develop schemas as they learn from experiences. People may also adjust their confidence, risk, and expectations, with respect to the cultural norms that they are accustomed to and the examples they see on a daily basis (Vitell et al., 1993). Stress, anxiety, and fatigue also diminishes performance (Croskerry, 2005).
Cognitive ability, culture, emotions, and feelings are all important contributors to a person’s readiness and ability to work through a strategy and achieve a goal. The individual decision maker must self-regulate, monitor, and adjust (if possible) these factors to create their optimal starting ground for decision performance, and ultimately achieve their goals.
Biases and Heuristics
For our ancestors, decision-making was often a matter of life or death. Heuristics and biases evolved as a way for the mind to process information quickly for near term survival (Buss, 2015). However, as major threats to survival have subsided and lifespans have increased, cognitive strategies that served well in the evolutionary past do not necessarily translate well to long-term goals. Many individuals today make poor decisions that prioritize immediate gratification. Although these heuristics minimize effort and maximize performance, these mental shortcuts may also lead to limited perspective, errors, and bias.
Minimizing Cognitive load
Humans are cognitive misers that utilize shortcuts in order to conserve mental energy (Fiske & Taylor, 2013). One strategy to minimize the onboarding of new information is confirmation bias, looking for and favoring information that support existing beliefs, thoughts, and mental models. Information may also be selectively omitted if it does not fit existing beliefs.
People also favor information that is recent and easily available. According to the recency bias, people recall information that has occurred recently better than they do older information (Arnold et al., 2000). Relatedly, individuals make judgments based on information that most easily comes to mind (Tversky & Kahneman, 1973). The availability bias has been supported in studies of everyday life. Ross and Sicoly’s (1979) study of married couples found that each partner showed egocentrism, overestimating their proportion of chores to be about 80%.
In uncertain situations, people are especially apt to use mental strategies to minimize mental effort. Attribute substitution is an automatic process by which a person substitutes a complex target attribute for a more easily computed heuristic attribute (Kahneman & Tversky, 1973). Trust, morals, stereotypes, emotions, and familiar experiences are particularly likely to be substituted in, due to their quick accessibility and familiarity. People may also reduce mental effort by following intuitive predictions. Kahneman and Tversky’s (1973) research also shows that both novices and experts tend to predict outcomes that are most representative, or typical of the evidence, regardless of the reliability of the evidence and knowledge of prior probabilities. Representativeness is also explained to contribute to unjustified confidence, as individuals erroneously select extreme values and rare events if they are representative of the evidence.
These biases are dangerous because they limit the scope of possible solutions and substantiate decisions and beliefs that may not be warranted. Design must thus account for these biases by minimizing user recall of extraneous and erroneous information, and providing clear pathways that highlight relevant information and validate this information in the user’s mind.
Efficiency
In order to increase efficiency, people also use strategies to filter through, simplify, and cope with the varied amounts of information they receive. One such strategy is eliminating individual aspects one by one, akin to a filter on a search engine, in which options that do not fit the criterion are eliminated (Tversky, 1972; Sternberg, 2012). Satisficing is another mental heuristic in line with bounded rationality. In satisficing, options are considered individually and the first option that meets a minimum threshold of acceptability is chosen (Sternberg, 2010; Simon, 1959). Satisficing is an effective decision making strategy when problems are complex and time is short. It allows a high return on investment and resources can then be placed on other decisions. However, for other decisions, like the high-risk situation of disease detection, a satisfactory solution may not suffice. Thus, people should consciously evaluate the context and the risks and adjust their strategies accordingly while making decisions.
Framing
Behavioral economics also suggests that people change their behavior based on how information is presented. Anchoring is a mental heuristic in which individuals adjust evaluations based on reference numbers. People given irrelevant primes may generate response estimates that are in line with primed information (Tversky & Kahneman, 1974).
Case: Coffee Meets Bagel
Dating is complicated. It is the connection of people who may have competing interests, beliefs, and emotions. In the past decade, online dating has experienced tremendous growth. Coffee Meets Bagel (CMB) is a dating app that employs several methods designed to make navigating complex dating decisions easier. Because the app is designed differently for men versus women, this review will focus on women’s experiences with the app and the basic goal of going on a date with a man. These decisions will be evaluated with respect to metacognition and recommendations for design will be provided.
Profile setup and preferences
CMB profile setup is quick, with fields for demographics, location, and personalized information. Women then set their preferences for prospective matches according to many of these filters- gender, age, height, distance, ethnicity, and religion. This filtering function can be related to the elimination of aspects, as men who do not fulfill criteria are not included in the match pool, and thus matches received are more in line with expectations and goals. This also streamlines the process of receiving matches that fulfill certain requirements, saving time and reducing risk. However, this “preference selection” can be further modified to align more closely to subjective interests. All these filters are equally weighted, but people may often consider those that satisfice, or fit most basic requirements. Potential dates may be lost as a result. Thus, one recommendation is for CMB to allow users to set weights of high, medium, and low for each preference. This way, the CMB matching algorithm will generate quality profiles that comprehensively account for individual preferences.
Selecting Matches
CMB’s platform is unique from other dating platforms because it is based on limited choice. Men receive 20 profiles per day to accept or decline, whereas women receive only five profiles of men that already like them. According to prospect theory, women benefit from this design, as they are certain that men have expressed interest, so there is more potential for a rewarding date and lower risk of rejection. The limited choice design forces picker women to think more carefully about each profile they receive and satisfice among the choices available, particularly for traits such as appearance, that are not selectable in preferences.
Self-Appraisal and Monitoring
A woman may gain a sense of how attractive she is to others based on her real life experiences, however this may not directly translate to online dating success. To help, CMB’s provides useful tools that facilitate self-appraisal and monitoring. Through the photo lab, users upload photos and receive feedback from other members on which photos are better to include in the profile. The tip page also provides the user advice on what to include and exclude from a profile following best practices. Finally, the “My Stats” page allows women to determine their relative levels success, through measures of pickiness (passes), inactivity (no connection), and productivity (connections). All these factors give a woman a good starting ground for dating success. Women are guided to modify their profile information and photos in accordance with guidelines and feedback, and then monitor the effectiveness of these strategic changes by reviewing their stats.
Furthermore, the “My Stats” tool may also be used to measure the effectiveness of women’s selection strategies. Women experiencing low connection numbers may choose to as modify their preferences and reduce pickiness in order to receive more connections.
Reducing Risk (Loss Aversion)
Although individuals may be matched, there is no guarantee that these individuals will go on the date, or even initiate conversation with one another. CMB, however, facilitates continued interactions between users by limiting the time frame for matches to chat. CMB also sends frequent notifications that remind users that chat time is running out (see screenshot). Humans are loss averse, so women who have already invested time in considering and liking their matches may be unwilling to lose them. This constraint may encourage women overcome any social or cultural misgivings and to initiate conversation.
Conclusion
As CMB shows, user decisions may be facilitated by metacognitive features that play on the human tendencies to self-evaluate, reduce risk, maximize gains, monitor progress, and adjust strategies and goals. However, it must be considered, are these guided decisions good? CMB assumes gender differences, which are exploited to nudge women towards further interaction. CMB should reevaluate this gender distinction, and provide women the male option of multiple swipes and the ability to proactively indicate interest in males.
Design may harness metacognitive tools to compensate for human’s limited cognitive bandwidth and resultant biases, by providing strategies and encouragement for the user to make structured decisions towards the achievement of a goal. However these metacognitive tools must be used ethically, balancing structured guidance with human agency.
Works Cited
Arnold, V., Collier, P. A., Leech, S. A., & Sutton, S. G. (2000). The effect of experience and complexity on order and recency bias in decision making by professional accountants. Accounting & Finance, 40(2), 109-134.
Buss, D. (2015). Evolutionary psychology: The new science of the mind. New York, New York: Psychology Press.
Croskerry, P. (2005). The theory and practice of clinical decision-making. Canadian Journal of Anesthesia/Journal canadien d'anesthésie, 52, R1-R8.
Edwards, W. (1954). The theory of decision making. Psychological Bulletin, 51, 380–417.
Fiske, S. T., & Taylor, S. E. (2013). Social cognition: From brains to culture. London, England: Sage. 15-17
Flavell, J. H. (1979). Metacognition and cognitive monitoring: A new area of cognitive–developmental inquiry. American psychologist, 34(10), 906.
Kahneman, D. (2013). Prospect Theory: An analysis of Decision under Risk. Journal of Chemical Information and Modeling, 53(9), 1689–1699.
Kahneman, D., & Frederick, S. (2002). Representativeness revisited: Attribute substitution in intuitive judgment. Heuristics and biases: The psychology of intuitive judgment, 49.
Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological review, 80(4), 237.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica: Journal of the econometric society, 263-291.
Kahneman, D., & Tversky, A. (1983). Choices, values, and frames. American psychologist, 39(4), 341.
Lerner, J. S., Li, Y., Valdesolo, P., & Kassam, K. S. (2015). Emotion and decision making. Annual Review of Psychology, 66, 799-823.
Locke, E. A., & Latham, G. P. (2006). New directions in goal-setting theory. Current directions in psychological science, 15(5), 265-268.
Najmaei, A., & Sadeghinejad, Z. (2016). An Integrative Review and a Research Agenda.
Nelson, T. O. (1990). Metamemory: A theoretical framework and new findings. Psychology of learning and motivation, 26, 125-173.
Paulson, E., & Bauer, L. (2011). Goal Setting as an Explicit Element of Metacognitive Reading and Study Strategies for College Reading. NADE Digest, 5(2009), 41–49.
Ross, M., & Sicoly, F. (1979). Egocentric biases in availability and attribution. Journal of personality and social psychology, 37(3), 322-336.
Shimamura, A. P. (2008). A neurocognitive approach to metacognitive monitoring and control. Handbook of Metamemory and Memory., 373–390. https://doi.org/10.4324/9780203805503.ch19
Simon, H. A. (1959). Theories of decision-making in economics and behavioral science. The American economic review, 49(3), 253-283.
Sternberg, R. J., & Sternberg, K. (2012). Cognitive psychology. Nelson Education.
Tversky, A. (1972). Elimination by aspects: A theory of choice. Psychological review, 79(4), 281.
Tversky, A., & Kahneman, D. (1974). Judgement under Uncertainty: Heuristics and Biases. Science, 185(4157), 1124–1131. https://doi.org/10.1126/science.185.4157.1124
Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453–458. https://doi.org/10.1126/science.7455683
Vitell, S. J., Nwachukwu, S. L., & Barnes, J. H. (1993). The Effects of Culture on Ethical Decision-Making : An Application of Hofstede ’ s Typology. Journal of Business Ethics, 12(1984), 753–760. https://doi.org/10.1007/BF00881307
Waters, H. S., & Schneider, W. (Eds.). (2010). Metacognition, strategy use, and instruction. New York, New York: Guilford Press.