Hard-easy Effect Research Articles

Using Behavioural Economics to Explain Substance Abuse Among Indian Adolescents

Understanding the behavioural factors that influence adolescent risky behaviour including substance abuse in India is critical for government agencies and organisations working to combat the now upsurging problem caused by substance use amongst adolescents. The objective of this study is to find out the behavioural biases that influence risky behaviour, including substance abuse among adolescents in India. An exploratory factor analysis is conducted on the responses of 682 teenage respondents from a close-ended questionnaire administered electronically in five schools in Ahmedabad, Gujarat. The Bandwagon Effect, Hard Easy Effect, Optimism Bias, Sunk Cost Fallacy, Social Norms, Disposition Effect, Illusion of Validity, and Availability Heuristic emerged as the factors most affecting the risky behaviour of adolescents including substance abuse. The findings of this research can be used to develop effective interventions to mitigate the effect of these biases on adolescent substance abuse and other risky behaviours. This study enriches the behavioural economics literature by focusing on teenage substance abuse, filling a critical research gap, and providing valuable insights for public health interventions. It also provides benchmarks for future research and policy-making in behavioural economics.

Unpacking Overconfident Behavior: Revisiting the Hard-Easy Effect

We investigate overconfident behavior as resulting from two discrepancies from Bayesian behavior: a bias in beliefs, usually known as overconfidence, and a bias in attitude towards ambiguity. We propose a method, valid under recent theories of decision under uncertainty, to separate the effects of subjective probabilities and ambiguity attitudes on overconfident behavior in two incentivized experiments. Our first study reveals the impact of ambiguity attitudes on overconfident behavior. While subjective probabilities were underconfident, subjects were ambiguity seeking. In the second study, our method replicates the hard-easy effect on beliefs: easy tasks lead to understimation of absolute performance and overplacement whereas hard tasks lead to the opposite. Additionally, we find that the task’s difficulty affects not only beliefs but also ambiguity attitudes. Easier tasks lead to more optimistic ambiguity attitudes for absolute and relative performance. In other words, while hard and easy tasks have opposite patterns on the belief component of overconfident behavior, easy tasks always contribute more to overconfident behavior than hard tasks.

The evolution of cognitive biases in human learning

Cognitive biases like underinference, the hard-easy effect, and recurrently non-monotonic confidence are evolutionarily puzzling when viewed as persistent flaws in how people learn from environmental feedback. To explain these empirically robust cognitive biases from an evolutionary perspective, we propose a model of ancestral human learning based on the cultural-evolutionary-theoretic hypothesis that the primary selection pressure acting on ancestral human cognition pertained not to learning individually from environmental feedback, but to socially learning task-specific knowledge. In our model—which is inspired by classical Bayesian models—an ancestral human learner (the student) attempts to learn task-specific knowledge from a role model, with the option of switching between different tasks and role models. Suppose that the student’s method of learning from their role model is a priori uncertain—in that it can either be successful imitation learning or de facto innovation learning—and the ecological fitness costs of meaningfully retaining environmental feedback are high. Then, the student’s fitness-maximizing strategy does not retain their environmental feedback and—depending on the choice of model parameters—can be characterized by all of the aforementioned cognitive biases. Specifically, in order for the evolutionarily optimal estimate of confidence in this learning environment to be recurrently non-monotonic, it is necessary (as long as the environment’s marginal payoff function satisfies a plausible quantitative condition) that a positive proportion of ancestral humans’ attempted imitation learning was unknowingly implemented as de facto innovation learning. Moreover, an ecologically rational strategy of selective social learning can plausibly cause the evolutionarily optimal estimate of confidence to be recurrently non-monotonic in the empirically documented way: general increase with an intermediate period of decrease.

Knowledge overconfidence among entrepreneurs from Austria and North Macedonia

Entrepreneurial decision-making is often characterized by unrealistic optimism. This so-called “overconfidence bias” has consequently received an increasing amount attention from the authors of literature on entrepreneurship research. Most empirical studies, however, that target the overconfidence bias have been conducted with students or samples from the general population. This study contributes to the entrepreneurial decision-making literature by explicitly targeting entrepreneurs from Austria and North Macedonia (n = 187), defined as individuals who have started their own businesses or are in the process of doing so. The entrepreneurs were asked to complete a general-knowledge questionnaire and to estimate the accuracy of their answers. The results suggest that the most of these entrepreneurs exhibit overconfidence (and some, underconfidence) with regard to the accuracy of their knowledge. Significant differences were observed between the bias scores and confidence scores of the Austrian and Macedonian respondents, but not the accuracy scores. The so-called “hard-easy effect,” which indicates that individuals display overconfidence when answering hard questions, but underconfidence when answering easy questions, was observed among the Austrian but not among the Macedonian entrepreneurs.

Bayesian inference with incomplete knowledge explains perceptual confidence and its deviations from accuracy

In perceptual decisions, subjects infer hidden states of the environment based on noisy sensory information. Here we show that both choice and its associated confidence are explained by a Bayesian framework based on partially observable Markov decision processes (POMDPs). We test our model on monkeys performing a direction-discrimination task with post-decision wagering, demonstrating that the model explains objective accuracy and predicts subjective confidence. Further, we show that the model replicates well-known discrepancies of confidence and accuracy, including the hard-easy effect, opposing effects of stimulus variability on confidence and accuracy, dependence of confidence ratings on simultaneous or sequential reports of choice and confidence, apparent difference between choice and confidence sensitivity, and seemingly disproportionate influence of choice-congruent evidence on confidence. These effects may not be signatures of sub-optimal inference or discrepant computational processes for choice and confidence. Rather, they arise in Bayesian inference with incomplete knowledge of the environment.

Item-level monitoring, response style stability, and the hard-easy effect

Cognitive and metacognitive processes during learning depend on accurate monitoring, this investigation examines the influence of immediate item-level knowledge of correct response feedback on cognition monitoring accuracy. In an optional end-of-course computer-based review lesson, participants (n = 68) were randomly assigned to groups to receive either immediate item-by-item feedback (IF) or no immediate feedback (NF). Item-by-item monitoring consisted of confidence self-reports. Two days later, participants completed a retention test (IF = NF, no significant difference). Monitoring accuracy during the review lesson was low, and contrary to expectations was significantly less with immediate feedback (IF < NF, Cohen’s d = .62). Descriptive data shows that (1) monitoring accuracy can be attributed to cues beyond actual item difficulty, (2) a hard-easy effect was observed where item difficulty was related to confidence judgements as a non-monotonic function, (3) response confidence was predicted by the Coh-Metrix dimension Word Concreteness in both the IF and NF treatments, and (4) significant autocorrelations (hysteresis) for confidence measures were observed for NF but not for IF. It seems likely that monitoring is based on multiple and sometimes competing cues, the salience of each relates in some degree to content difficulty, but that the stability of individual response styles plays a substantive role in monitoring. This investigation shows the need for new applications of technology for monitoring multiple measures on the fly to better understand SRL processes to support all learners.

Exploring the Effects of Overconfidence Bias and Hard-easy Effect in Self-monitoring of Biological Concept Test

Exploring the Effects of Overconfidence Bias and Hard-easy Effect in Self-monitoring of Biological Concept Test

The Evolution of Cognitive Biases in Human Learning

Cognitive biases like underinference, the hard-easy effect, and recurrently non-monotonic confidence are evolutionarily puzzling when viewed as persistent flaws in how people learn from environmental feedback. To explain these empirically robust cognitive biases from an evolutionary perspective, we propose a model of ancestral human learning based on the cultural-evolutionary-theoretic hypothesis that the primary selection pressure acting on ancestral human cognition pertained not to learning individually from environmental feedback, but to socially learning task-specific knowledge. In our model—which is inspired by classical Bayesian models—an ancestral human learner (the student) attempts to learn task-specific knowledge from a role model, with the option of switching between different tasks and role models. Suppose that the student's method of learning from their role model is a priori uncertain—in that it can either be successful imitation learning or de facto innovation learning—and the ecological fitness costs of meaningfully retaining environmental feedback are high. Then, the student's fitness-maximizing strategy does not retain their environmental feedback and—depending on the choice of model parameters—can be characterized by all of the aforementioned cognitive biases. Specifically, in order for the evolutionarily optimal estimate of confidence in this learning environment to be recurrently non-monotonic, it is necessary (as long as the environment's marginal payoff function satisfies a plausible quantitative condition) that a positive proportion of ancestral humans' attempted imitation learning was unknowingly implemented as de facto innovation learning. Moreover, an ecologically rational strategy of selective social learning can plausibly cause the evolutionarily optimal estimate of confidence to be recurrently non-monotonic in the empirically documented way: general increase with an intermediate period of decrease.

Does Performance Calibration Generalize Across Sport Tasks? A Multiexperiment Study in Physical Education.

Three experiments examined students' calibration in physical education in relation to task characteristics. Participants in the 3 experiments were 388 students. Calibration accuracy and bias were calculated based on students' predicted and actual performance in tests including variations of a sport task (basketball shooting) and tasks from different sports (basketball and soccer). An overconfidence effect was found in all experiments, and evidence regarding the hard-easy effect emerged. High compared with low performers were more accurate, and some variations with respect to gender also emerged. The magnitude of calibration error was similar across tasks, whereas approximately half of the students were consistent in the direction of calibration (most of them were overestimators). Results are discussed with reference to theoretical and empirical evidence associated with performance calibration and self-regulated learning in physical education. Methodological issues, practical implications, and future directions are also discussed.

Inaccuracy and Overconfidence in Metacognitive Monitoring of University Students

The aim of the paper is to stress the relationship between individual metacognitive accuracy and academic performance. Moreover, we tested the relationship between the accuracy, items difficulty and bias score and exam results. Metacognitive monitoring calibration of 100 university students was tested in exam settings, using postdicted confidence judgments. Absolute local and global accuracy and total bias score were related with test performance, difficulty and types of the items: multiple choice and open ended items. The most important results show local and local inaccuracy or overconfidence, but also an unexpected greater accuracy on low performing subjects compared with there’s more performing counterparts. The open ended items low, but positively correlate with metacognitive monitoring inaccuracy, both local and global, a possible illustration of the hard-easy effect. The bias score is globally negatively related with performance, but the more the subjects respond to the difficult open-ended items, the lower the bias in self-appreciation. We conclude that there is a global expected relationship between test results and both accuracy and bias score. Also, particular results show a more nuanced relationship between local and global accuracy, items difficulty and type and bias score. Some theoretical and methodological issues are discussed and futures research direction is proposed.

Overconfidence and forecast accuracy

Purpose This paper aims to investigate how overconfidence bias affects financial market participants’ forecast accuracy based on the hard–easy effect concept of overconfidence research. Design/methodology/approach The authors adopt an experimental method for behavioural finance studies. In the experiment, the authors measure and capture participants’ forecast accuracy as well as their individual confidence level. In particular, participants make incentive-compatible forecasts, that is, the elicited forecast determines the participants’ financial rewards in real monetary gain/loss. Findings The results show that the hard–easy effect causes optimistic forecasts for hard-to-predict stocks, indicating that overconfident investors tend to make over-optimistic and less accurate price forecasts when making judgements on hard tasks. Consistent with the literature, the authors find evidence of a negative relationship between forecast accuracy and confidence level. The results also indicate that the overall relationship between overconfidence and forecast accuracy is driven by the price forecasts made for hard-to-predict stocks. Originality/value As per the authors’ knowledge, this paper is one of the first studies that provides empirical evidence directly showing the hard–easy effect in the relation between overconfidence and forecast ability in an experimental setting. This study uses an experimental design that specifically measures the hard–easy effect in a stock market scenario using professional financial information and real monetary incentives, which have not been used in any previous studies of the hard–easy effect.

Examining high-school students’ overconfidence bias in biology exam: a focus on the effects of country and gender

ABSTRACTAccurate, rational, and scientific decision making is now considered to be the most important skill in science education. Many studies have found that overconfidence bias is one of the cognitive biases hindering people from achieving such decision making. Gender and country play crucial roles in overconfidence bias. For instance, some particular cultures and genders tend to be more overconfident than others. However, whether or not the two variables interact to influence overconfidence bias also indirectly influences decision making, especially in the context of science education. The purpose of this study is to identify the effects of country and gender on performance, confidence, and overconfidence bias in the samples of Indonesian and Korean high-school students while doing on a biology exam. The twenty-one American Association for the Advancement of Science (AAAS) questions on the topics of genetics and evolution were administered to 297 Indonesian and 235 Korean high-school students, in their first and second years. Every question was featured with a question asking students how confident they are in answering the question correctly. The two-way Analysis of Variances (2-way ANOVA) test was used to answer the research questions. Based on the analyses, we found no significance interactional effects of gender and country in test scores. In contrast, we found a significant interactional effects in both confidence in genetics and evolution. Regarding overconfidence bias, for which that we merged both concepts, we found that country had a higher influence on students’ overconfidence bias than did gender. Additionally, we found the hard-easy effect phenomenon followed overconfidence bias phenomenon. The relationships between country, gender, science education, cognitive bias, and overconfidence bias are discussed. Suggestions for reducing overconfidence bias are also provided.

Confidence biases and learning among intuitive Bayesians

We design a double-or-quits game to compare the speed of learning one’s specific ability with the speed of rising confidence as the task gets increasingly difficult. We find that people on average learn to be overconfident faster than they learn their true ability and we present an intuitive-Bayesian model of confidence which integrates confidence biases and learning. Uncertainty about one’s true ability to perform a task in isolation can be responsible for large and stable confidence biases, namely limited discrimination, the hard–easy effect, the Dunning–Kruger effect, conservative learning from experience and the overprecision phenomenon (without underprecision) if subjects act as Bayesian learners who rely only on sequentially perceived performance cues and contrarian illusory signals induced by doubt. Moreover, these biases are likely to persist since the Bayesian aggregation of past information consolidates the accumulation of errors and the perception of contrarian illusory signals generates conservatism and under-reaction to events. Taken together, these two features may explain why intuitive Bayesians make systematically wrong predictions of their own performance.

Judgment Extremity and Accuracy Under Epistemic vs. Aleatory Uncertainty

People view uncertain events as knowable in principle (epistemic uncertainty), as fundamentally random (aleatory uncertainty), or as some mixture of the two. We show that people make more extreme probability judgments (i.e., closer to 0 or 1) for events they view as entailing more epistemic uncertainty and less aleatory uncertainty. We demonstrate this pattern in a domain where there is agreement concerning the balance of evidence (pairings of teams according to their seed in a basketball tournament) but individual differences in the perception of the epistemicness/aleatoriness of that domain (Study 1), across a range of domains that vary in their perceived epistemicness/aleatoriness (Study 2), in a single judgment task for which we only vary the degree of randomness with which events are selected (Study 3), and when we prime participants to see events as more epistemic or aleatory (Study 4). Decomposition of accuracy scores suggests that the greater judgment extremity of more epistemic events can manifest itself as a trade-off between enhanced resolution and diminished calibration. We further relate our findings to the hard–easy effect and also show that differences between epistemic and aleatory judgment are amplified when judges have more knowledge concerning relevant events. Data, as supplemental material, are available at http://dx.doi.org/10.1287/mnsc.2015.2344 . This paper was accepted by Yuval Rottenstreich, judgment and decision making.

Category learning judgments in the classroom: Can students judge how well they know course topics?

A category learning judgment (CLJ) involves judging one’s learning or performance for a given topic or category. The present study was the first to investigate CLJs in a classroom, where students’ judgments of how well they have learned topics may be particularly relevant for guiding their study decisions. In an undergraduate statistics class, students predicted their performance on six different exam topics, as well as predicting their global exam performance, for each exam during the semester. Regarding the absolute accuracy of CLJs, we observed slight overestimation (bias), substantial deviation from accuracy (absolute bias), and little improvement across exams. Students’ CLJs varied among topics, but they were less variable than actual topic performance and were poor at discriminating well-learned from poorly-learned topics (i.e., low relative accuracy). We examined two factors predictive of CLJ accuracy: topic difficulty and student mastery of the topics. Regarding topic difficulty, a hard-easy effect was observed, such that more difficult topics produced greater overestimation and easier topics produced more underestimation. A hard-easy effect also extended to absolute bias: difficult topics produced larger deviations from accuracy than easy topics did. Regarding student mastery of topics, we found that lower mastery predicted CLJ overestimation and higher mastery predicted CLJ underestimation. Lower mastery was also associated with larger absolute bias. Compared to global judgments, CLJs were less accurate, although students were more confident in their CLJs. In sum, developing methods to improve the accuracy of CLJs in classrooms is an important direction for future research.

How do risk attitudes affect measured confidence?

We examine the relationship between confidence in own absolute performance and risk attitudes using two confidence elicitation procedures: self-reported (non-incentivised) confidence and an incentivised procedure that elicits the certainty equivalent of a bet based on performance. The former procedure reproduces the “hard-easy effect” (underconfidence in easy tasks and overconfidence in hard tasks) found in a large number of studies using non-incentivised self-reports. The latter procedure produces general underconfidence, which is significantly reduced, but not eliminated when we filter out the effects of risk attitudes. Finally, we find that self-reported confidence correlates significantly with features of individual risk attitudes including parameters of individual probability weighting.

Confidence estimation as a stochastic process in a neurodynamical system of decision making.

Evaluation of confidence about one's knowledge is key to the brain's ability to monitor cognition. To investigate the neural mechanism of confidence assessment, we examined a biologically realistic spiking network model and found that it reproduced salient behavioral observations and single-neuron activity data from a monkey experiment designed to study confidence about a decision under uncertainty. Interestingly, the model predicts that changes of mind can occur in a mnemonic delay when confidence is low; the probability of changes of mind increases (decreases) with task difficulty in correct (error) trials. Furthermore, a so-called "hard-easy effect" observed in humans naturally emerges, i.e., behavior shows underconfidence (underestimation of correct rate) for easy or moderately difficult tasks and overconfidence (overestimation of correct rate) for very difficult tasks. Importantly, in the model, confidence is computed using a simple neural signal in individual trials, without explicit representation of probability functions. Therefore, even a concept of metacognition can be explained by sampling a stochastic neural activity pattern.

Relation between belief and performance in perceptual decision making.

In an uncertain and ambiguous world, effective decision making requires that subjects form and maintain a belief about the correctness of their choices, a process called meta-cognition. Prediction of future outcomes and self-monitoring are only effective if belief closely matches behavioral performance. Equality between belief and performance is also critical for experimentalists to gain insight into the subjects' belief by simply measuring their performance. Assuming that the decision maker holds the correct model of the world, one might indeed expect that belief and performance should go hand in hand. Unfortunately, we show here that this is rarely the case when performance is defined as the percentage of correct responses for a fixed stimulus, a standard definition in psychophysics. In this case, belief equals performance only for a very narrow family of tasks, whereas in others they will only be very weakly correlated. As we will see it is possible to restore this equality in specific circumstances but this remedy is only effective for a decision-maker, not for an experimenter. We furthermore show that belief and performance do not match when conditioned on task difficulty – as is common practice when plotting the psychometric curve – highlighting common pitfalls in previous neuroscience work. Finally, we demonstrate that miscalibration and the hard-easy effect observed in humans' and other animals' certainty judgments could be explained by a mismatch between the experimenter's and decision maker's expected distribution of task difficulties. These results have important implications for experimental design and are of relevance for theories that aim to unravel the nature of meta-cognition.

A Target-Based Foundation for the 'Hard-Easy Effect' Bias

The effect is a well-known cognitive bias on self-confidence calibration that refers to a tendency to overestimate the probability of success in hard-perceived tasks, and to underestimate it in easy-perceived tasks. This paper provides a target-based foundation for this effect, and predicts its occurrence in the expected utility framework when utility functions are S-shaped and asymmetrically tailed. First, we introduce a definition of hard-perceived and easy-perceived task based on the mismatch between an uncertain target to meet and a suitably symmetric reference point. Second, switching from a target-based language to a utility-based language, we show how this maps to an equivalence between the hard-perceived target/gain seeking and the easy-perceived target/loss aversion. Third, we characterize the agent's miscalibration in self-confidence. Finally, we derive sufficient conditions for the Ohard-easy effectO and the reversed hard-easy effect to hold.

Toward a synthesis of cognitive biases: How noisy information processing can bias human decision making.

A single coherent framework is proposed to synthesize long-standing research on 8 seemingly unrelated cognitive decision-making biases. During the past 6 decades, hundreds of empirical studies have resulted in a variety of rules of thumb that specify how humans systematically deviate from what is normatively expected from their decisions. Several complementary generative mechanisms have been proposed to explain those cognitive biases. Here it is suggested that (at least) 8 of these empirically detected decision-making biases can be produced by simply assuming noisy deviations in the memory-based information processes that convert objective evidence (observations) into subjective estimates (decisions). An integrative framework is presented to show how similar noise-based mechanisms can lead to conservatism, the Bayesian likelihood bias, illusory correlations, biased self-other placement, subadditivity, exaggerated expectation, the confidence bias, and the hard-easy effect. Analytical tools from information theory are used to explore the nature and limitations that characterize such information processes for binary and multiary decision-making exercises. The ensuing synthesis offers formal mathematical definitions of the biases and their underlying generative mechanism, which permits a consolidated analysis of how they are related. This synthesis contributes to the larger goal of creating a coherent picture that explains the relations among the myriad of seemingly unrelated biases and their potential psychological generative mechanisms. Limitations and research questions are discussed.