Loading...

Heuristics: a source of judgement fallacies or decision-making aids?

Research Paper (undergraduate) 2012 30 Pages

Psychology - Methods

Excerpt

Table of Contents

ABSTRACT

1. INTRODUCTION

2. APPROACHES TO RATIONALITY

3. THE HEURISTICS-AND-BIASES FRAMEWORK
3.1. THE TWO-SYSTEMS VIEW
3.2. ACCESSIBILITY AND THE AVAILABILITY HEURISTIC
3.3. NEW HEURISTICS
3.3.1. A generic heuristic process
3.3.2. Extension of heuristics
3.3.3 How can System 2 override intuitive judgement (System 1)?
3.4. CRITIQUES TO HAB

4. FAST-AND-FRUGAL HEURISTICS
4.1. SIMILARITIES TO HAB
4.2. DIFFERENCES TO HAB
4.2.1. Ecological Rationality
4.2.2. Less-is-more
4.2.3. Bias-Variance Dilemma
4.2.4. A predictable model vs. labels
4.2.5. Adaptive Toolbox: Examples of fast-and-frugal heuristics
4.3. CRITIQUES
4.3.1. HAB’s reply
4.3.2. Questioning the predictive power of FAF
4.3.3. Evans and Over (2010)

5. CONCLUSION
5.1. ALTERNATIVE EXPLANATIONS AND LIMITS OF EXISTING RESEARCH
5.2. FUTURE DIRECTIONS: A POSSIBLE SYNTHESIS?

6. REFERENCES

Abstract

This review essay explores different perspectives and conceptualizations of the study of heuristics, decision-making rules which operate under constrained time and computation (Kahneman, 2011). Two opposed models of heuristics that assume conditions of bounded rationality, the heuristics-and-biases and the fast- and-frugal framework, are assessed. Whereas the former evaluates heuristics in terms of logical rationality and postulates that humans exhibit predictable fallacies in judgement, the latter focuses on ecological validity, and suggests that humans possess an adaptive toolbox of evolutionary developed decision-making rules which enable better decision making. Finally, alternative explanations and limitations of existing research programs will be explored, concluding with a demand for a rigorous evaluation of experimental designs as well as outlining conditions for a possible synthesis.

1. Introduction

Since the 1950s, psychologists have argued that humans are incapable of behaving rationally and optimal. Most prominently, Herbert Simon, advocates:

‘THE THEORY OF RATIONAL (‘SENSIBLE’) HUMAN BEHAVIOUR HAS BROKEN LOOSE FROM THE ILLUSORY AND EMPIRICALLY UNSUPPORTED NOTION THAT DECIDING RATIONALLY MEANS MAXIMIZING EXPECTED UTILITY.’ Herbert A. Simon, cited in Gigerenzer et al (1999)

In the last three decades, the study of bounded rationality has expanded greatly, especially with Kahneman and Tversky’s (1974, 1983) research program. Heuristics, of Greek origin (‘serving to find out or discover’), have been defined as intuitive ‘rules of thumb’ or simple strategies to deal with situations where resources are sparse, which all humans are thought to use similarly (Kahneman et al, 1982).

The present review aims to highlight the origins and definitions of heuristics on the basis of the two most dominant models, the heuristics-and- biases (HAB in the following) approach and the fast-and-frugal (FAF in the following) framework. Starting with underlying epistemological and philosophical assumptions of both approaches, and then continuously moving towards a comparative evaluation of each research group’s main arguments, this essay intends to highlight similarities, differences and possible criticisms. In the conclusion, both theoretical frameworks will be conceptualized, and possible integration possibilities as well as a lack of such an integration to date, will be discussed.

2. Approaches to rationality

In order to discuss the nature of heuristics, it is necessary to highlight different models and assumptions of rationality, which underlie the understanding of human decision-making (Marewski et al, 2010).

Models of unbounded rationality ask how people would behave if they were omniscient (capacity to know everything infinitely) and omnipotent (unlimited power) and assume that decision makers act as if they collect and evaluate all information to reach the mathematically optimal solution. This view is common in the studies of economics and assumes that human decision-making is infinitely flexible and domain-general (Payne et al, 1993).

Models of bounded rationality revise the assumption of rational decisionmakers, accounting for the fact that perfectly rational decisions are often not feasible in practice due to finite computational resources available for making them (Simon, 1956). The process of ‘satisficing’, seeking what is satisfactory rather than optimal, illustrates where behaviour deviates from the tenets of unbounded rationality (Simon, 1957). Where resources are sparse, individuals may employ simple strategies such as heuristics (Simon, 1978). In the following, two approaches to bounded rationality will be discussed.

3. The heuristics-and-biases framework

Since their ground-breaking publication in ‘Nature’, Kahneman and Tversky (1974) advocated the view that individuals rely on a limited number of heuristic principles when engaging in complex tasks. Although useful by reducing task complexity, these sometimes lead to severe and systematic errors.

In order to evaluate HAB, main assumptions and arguments of its’ underlying framework will be examined, before introducing other, more critical perspectives.

3.1. The two-systems view

Kahneman (2002) proposed that intuitive judgements occupy a position between the automatic perception and deliberate operations of reasoning. This distinction between intuition and reasoning, two types of cognitive processes, has been labelled System 1 and 2 (Stanovich & West, 2000).

It is hypothesized that System 1 operates automatically, quickly, with little effort, and no sense of voluntary control (Evans & Over, 2010). Similar to the features of perceptual processes, operations are associative, relatively flexible and potentially rule-governed (Petty & Cacioppo, 1986). A picture of an angry woman, for example, automatically evokes certain predispositions about the future (see Figure 1), such as the impression she may say unkind words in a loud voice (Kahneman, 2011).

Figure 1: A picture of an angry woman (adapted from Kahneman, 2011)

Abbildung in dieser Leseprobe nicht enthalten

System 2, on the other hand, deliberately processes information and allocates attention to effortful mental activities. It is involved in all judgements, whether they originate in impressions of System 1 or deliberate reasoning. Moreover, it monitors the quality of both mental operations and overt behaviour (Gilbert, 2002). Therefore, when presented with a multiplication problem (see Figure 2), a precise solution does not easily come to mind. In order to solve this equation, System 2 must be engaged in an effortful and slower thought process.

Figure 2: A multiplication problem (adapted from Kahneman, 2011)

Abbildung in dieser Leseprobe nicht enthalten

Intuitive answers to judgemental problems proposed by System 1 are monitored by System 2, which makes the explicit judgement and, through monitoring of these proposals, may endorse, correct or override them. However, monitoring is usually quite lax and allows the expression of many intuitive judgements, including some erroneous one’s. In one exemplary experiment, Kahneman (2003) posed the following question:

A BAT AND A BALL COST $1.10 IN TOTAL. THE BAT COSTS $1 MORE THAN THE BALL. HOW MUCH DOES THE BALL COST?

Over 50% of participants (students at Princeton University) responded erroneously (0.10$ instead of 0.05$), which illustrates how lightly the output of system 1 (which proposed the wrong answer of 0.10$) is monitored by system 2 (which, had it been utilized, would have calculated the correct answer). Moreover, complex cognitive operations can eventually migrate from system 2 to system 1 as proficiency and skill are acquired (Kahneman & Frederick, 2002), for example the ability of chess masters to instantly perceive the strength or weaknesses of chess positions.

3.2. Accessibility and the availability heuristic

According to HAB, judgements of frequency are made on the basis of item accessibility, a System 1 heuristic, which is replaced by a focus on content when System 2 is more engaged. The ease with which particular mental concepts come to mind (Higgins, 1996) operates on the notion that ‘if you can think of it, it must be important’ (Esgate & Groom, 2004). To comprehend the workings of System 1, it is important to understand why some thoughts are more accessible than others. For example, the average length of lines in Figure 3 is more easily accessible than the total length of lines.

Abbildung in dieser Leseprobe nicht enthalten

Figure 3: vertical lines of different lengths (adapted from Kahneman, 2002)

Consequently, some attributes seem to be more accessible. These socalled ‘natural assessments’ are routinely and automatically registered by the perceptual system without intention or effort. Conversely, accessibility is by no means representative of probability, and so heuristics lead to biases and erroneous judgements (Tversky & Kahneman, 1983).

3.3. New Heuristics

Tversky and Kahneman’s original theory (1974, 1983) included three heuristics of judgement: representativeness (judgements influenced by what is typical), availability (judgements based on what comes easily to mind) and anchoring (judgements relying on what comes first). The updated model of heuristics proposed by Kahneman and Frederick (2002) departs from the original formulation in three ways:

1. A common process of attribute substitution explains how heuristics work.
2. Extension of the concept of heuristics beyond the domain of judgements about uncertain events.
3. An explicit treatment of the conditions under which intuitive judgements will be modified or overridden by the monitoring operations associated with System 2.

3.3.1. A generic heuristic process

Attribute substitution describes the process of substitution of a specific attribute of a judged object with a related one that comes to mind more easily (i.e. is more readily accessible) as a means to reduce effort. Therefore, when confronted with a difficult question, an easier one may be answered instead outside of conscious awareness (Gilovic et al, 2002). Because the target attribute and the heuristic attribute are different, the substitution of one for the other inevitably introduces systematic biases, where humans violate a law of logic, probability or another standard of rationality. Hence, respondents to a difficult problem offer a reasonable answer to a question they have not been asked. For example, a survey of German students included the following two questions (Kahneman, 2011):

[...]

Details

Pages
30
Year
2012
ISBN (eBook)
9783656346036
ISBN (Book)
9783656346425
File size
652 KB
Language
English
Catalog Number
v203666
Institution / College
University of St Andrews
Grade
20
Tags
Heuristics Psychology Decision-Making Cognitive Cognition Kahneman Gigerenzer

Author

Previous

Title: Heuristics: a source of judgement fallacies or decision-making aids?