Helen is angry with her husband Lewis who avoids approaching his boss for a pay rise Lewis argues that the timing is not right. Helen says he simply falls to face up to people. How are these attributions different in kind? You read a newspaper report about a rape case in which the defence lawyer pointed out that, the young woman who was the victim was dressed provocatively. What attributional error is involved here? ‘The job market was tight and Rajna began to worry that she might be made redundant, Then she heard a rumour that the worst had come - several staff were about to be fired. She was itching to pass this on to the next colleague that she saw. Why would Rajna want to spread the rumour further?84 CHAPTER 3 ATTRIBUTION AND SOCIAL EXPLANATION. Axtribution| ‘The proces of assigning = «cause our own behaviour and that of others, In search of the meaning of life Religion isan expression of a most fundamental need to understand our world, Like all novices, young Buddhists have much to lear, Seeking the causes of behaviour People are preoccupied with finding, constructing and testing explanations of their experi- ences, We try to understand our world to make it orderly and meaningful enough for adap- tive action, and we feel uncomfortable if we do not have such an understanding, In particulas, ‘we need to understand people. Through life most of us construct adequate explanations (i. theories) of why people behave in certain ways; in this respect, we are all ‘naive’ or lay psy- chologists. This is extraordinarily useful, because it allows us (with varying accuracy) to predict how someone will behave, and possibly to influence whether someone will behave in that way or not. Thus, we gain some control over our destiny. People construct explanations for both physical phenomena (e.g. earthquakes, the sea- sons} and human behaviour (e.g. anges, a particular attitude}, and in general such explana- tions are causal explanations, in which specific conditions are attributed a causal role. Causal explanations are particularly powerful bases for prediction and control (Hilton, 2007). In this chapter, we discuss how people make inferences about the causes of their own and other people’s behaviour, and the antecedents and consequences of such inferences. Social psychological theories of causal inference are called attribution theories (Hewstone, 1989; Ross & Fletcher, 1985; Smith, 1994; Trope & Gaunt, 2007; Weary, Stanley, & Harvey, 1989) ‘There are seven main theoretical emphases that make up the general body of attribution theory: 1 Heider’s (1958) theory of naive psychology; 2 Jones and Davis's (1965) theory of correspondent inference; 3 Kelley's (1967) covariation model; 4 Schachter’s (1964) theory of emotional lability; 5. Bem’s (1967, 1972] theory of self-perception;HOW PEOPLE ATTRIBUTE CAUSALITY 85 6 Weiner’s (1979, 1985) attributional theory; and 7 Deschamps's (1983), Hewstone’s (1989) and Jaspars’s (Hewstone & Jaspars, 1982, 1984) intergroup perspective, ‘We discuss the first six of these below and then deal with intergroup attribution by itself in greater detail later in the chapter. How people attribute causality People as naive psychologists Fritz Heider (1958) believed it was important for social psychologists to study people’s naive, or common sense, psychological theories, because such theories influenced ordinary peo- ple’s everyday perceptions and behaviour. For example, people who believe in astrology are likely to have different expectations and to act in different ways from those who do not. Heider believed that people are intuitive psychologists who construct causal theories of human behaviour, and because such theories have the same form as scientific social psycho- logical theories, people are actually intuitive or naive psychologists. Heider based his ideas on three principles: 1 Because we feel that our own behaviour is motivated rather than random, we look for the causes for other people’s behaviour in order to discover their motives. The search for causes does seem to pervade human thought, and it can be difficult to explain or com- ment on something without using causal language. Heider and Simmel (1944) demon- strated this in an ingenious experiment. People who were asked to describe the movement of abstract geometric figures described them as if they were humans with intentions to act in certain ways. Nowadays, we can witness the same phenomenon in people’s often highly emotional ascription of human motives to computer-generated figures. People’s pervasive need for causal explanation reveals itself most powerfully in the way that almost all societies construct an origin myth, an elaborate causal explanation for the origin and meaning of life that is often a centrepiece of a religion, 2 Because we construct causal theories in order to be able to predict and control the envi- ronment, we tend to look for stable and enduring properties of the world around us. We try to discover personality traits and enduring abilities in people, or stable properties of situations, that cause behaviour. 3 In attributing causality for behaviour, we distinguish between personal factors (e.g. per sonality, ability) and environmental factors (e.g. situations, social pressure). The former are examples of an internal (or dispositional) attribution and the latter of an external (or situational) attribution. So, for example, it might be useful to know whether some- one you mect at a party who seems aloof and distant is an aloof and distant person or is acting like that because that person is not enjoying the party. Heider believed that because internal causes, or intentions, are hidden from us, we can only infer their pres- cence if there are no clear external causes. However, as we see later, people tend to be biased in preferring internal to external attributions even in the face of evidence for external causality. It seems that we readily attribute behaviour to stable properties of people. Scherer (1978), for example, found that people made assumptions about the stable personality traits of complete strangers simply on the basis of hearing their voices on the telephone. Heider identified the major themes and provided the insight that forms the blueprint for all subsequent, more formalised, theories of attribution, Naive psychologist (erscientist) Model of socal cagntion that characterises people sing tonal science caustvetect analyses to understand reir wore Internal (or dispositional) attribution Proce of asgning the ‘couse of our cin or other? benaviourto internal or ‘spostona factors External (or stuatonal) attribution Assigning he cause of our fn or other behaviour feeroa or envtonmentl factors86 CHAPTER3 ATTRIBUTION AND SOCIAL EXPLANATION CCorrespondentinference ‘Causal atibuton of behaviour to underyng depositions Non-common effects fects of benviou that ae revel eclsvet that behaviour rainertnan other behaviours ‘Outcome bias Belt hat the cutcomes of behaviour were menced bythe person wha cnase ‘yebehaviow Figure 3.1 How we make a correspondent inference From acts to dispositions Ned Jones and Keith Davis's (1965; Jones & McGillis, 1976) theory of correspondent inference explains how people infer that a person's behaviour corresponds to an underlying disposition or personality trait —how we infer, for example, that a friendly action is due to an underlying disposition to be friendly. People like to make correspondent inferences (attribute behaviour to underlying disposition) because a dispositional cause isa stable cause that makes people’s behaviour predictable and thus increases our own sense of control over our world. To make a correspondent inference, we draw on five sources of information, of cues (see Figure 3.1) 1 Freely chosen behaviour is more indicative of a disposition than is behaviour that is clearly under the control of external threats, inducements or constraints. 2. Behaviour with effects that are relatively exclusive to that behaviour rather than common, to many behaviours (ie. behaviour with non-common effects) tells us more about dispo- sitions. People assume that others are aware of non-common effects and that the specific behaviour was performed intentionally to produce the non-common effect ~this tendency has been called outcome bias (Allison, Mackie, & Messick, 1996). So, for example, if a person has to choose between behaviour A and behaviour B, and both produce roughly the same effects {i.e. no non-common effects) or a very large number of different effects (ie. many non-common effects), the choice tells us little about the person's disposition. However, if the behaviours produce a small number of different effects (ie. few non- common effects ~ e.g, behaviour A produces only terror and behaviour B produces only joy}, then the choice does tell us something about that person’s disposition. 3. Socially desirable behaviour tells us little about a person’s disposition, because itis likely to be controlled by societal norms. However, socially undesirable behaviour is generally counter-normative and is thus a better basis for making a correspondent inference. Correspondent inference ‘The act reflects some “true characteristic of the person (trait, motive, intention, attitude, etc) ‘To make an inference that a person's behaviour corresponds to an underlying ispostion, we draw on five sources of informationHOW PEOPLE ATTRIBUTE CAUSALITY 87 4 We make more confident correspondent inferences about others’ behaviour that has important consequences for ourselves: that is, behaviour that has hedonic relevance. 5. We make more confident correspondent inferences about others’ behaviour that seems to be directly intended to benefit or harm us: that is, behaviour that is high in personalism, Experiments testing correspondent inference theory provide some support. Jones and Harris (1967) found that American students making attributions for speeches made by other students tended to make more correspondent inferences for freely chosen socially unpopular positions, such as freely choosing to make a speech in support of Cuba’s president at the time, Fidel Castro. In another experiment, Jones, Davis and Gergen (1961) found that participants made more correspondent inferences for out-of-role behaviour, such as friendly, outer-directed behaviour by someone who was applying for an astronaut job, in which the required attri utes favour a quiet, reserved, inner-directed person. Correspondent inference theory has some limitations and has declined in importance as an attribution theory (Hewstone, 1989; Howard, 1985). For instance, the theory holds that correspondent inferences depend significantly on the attribution of intentionality, yet unin- tentional behaviour (e.g. careless behaviour) can be a strong basis for a correspondent infer- ence (c.g, that the person is a careless person). ‘There is also a problem with the notion of non-common effects. Correspondent inference theory maintains that people assess the commonality of effects by comparing chosen and non-chosen actions, while research shows that people simply do not attend to non-oceurring behaviours and so would not be able to compute the commonality of effects accurately (Nisbett & Ross, 1980; Ross, 1977). More generally, although we may correct dispositional attributions in the light of situational factors, this is a rather deliberate process, whereas correspondent inferences themselves are relatively automatic (Gilbert, 1995). People as everyday scientists ‘The best-known attribution theory is Harold Kelley’s (1967, 1973) covariation model. In try- ing to discover the causes of behaviour, people act much like scientists. They identify what factor covaries most closely with the behaviour and then assign to that factor a causal role. The procedure is similar to that embodied by the statistical technique of analysis of variance (ANOVA), and for this reason Kelley’s model is often referred to as an ANOVA model People use this covariation principle to decide whether to attribute behaviour to internal dis- positions (e.g, personality) or external environmental factors (e.g. social pressure) In order to make this decision, people assess three classes of information associated with the co-occurrence of a certain action (¢.g. laughter) by a specific person (e.g. Tom) with a potential cause (e.g, a comedian): 1 Consistency information — does Tom always laugh at this comedian (high consistency) or ‘only sometimes laughs at this comedian {low consistency]? 2. Distinctiveness information ~does Tom laugh at everything (low distinctiveness) or only at the comedian (high distinctiveness)? 3. Consensus information ~ does everyone laugh at the comedian (high consensus) or is it only Tom who laughs (low consensus)? ‘Where consistency is low, people discount the potential cause and search for an alterna- tive (see Figure 3.2). If Tom sometimes laughs and sometimes does not laugh at the come- dian, then presumably the cause of the laughter is neither the comedian nor Tom but some other covarying factor: for example, whether or not Tom smoked marijuana before listening to the comedian, or whether or not the comedian told a funny joke (see McClure, 1998, for a review of the conditions under which discounting is most likely to occur), Where Hedonic relevance Retest behaviour that has Inport cee concequences for set Personalism Benaviout int appears to be iret intendec to benef ar harm onesei rather than others Covariation model Keleys teary of causa sttouton = people eign the cause of behaviour to the facto: that covaries most cose wth the behaviour Consistency information Information about he exert to which a behaviour slays covoceurs th 3 stimulus x Distinciveness information Iivormation about whether a persons reacion occu's only with one stimulus oris Consensus information Information about he econ towhen ather people react nye sme vray t0 a status x Discount relsionship betieen a {pac case and a pectic behaviour that ease i counted in favour of ome other ease88 CHAPTER 3 ATTRIBUTION AND SOCIAL EXPLANATION. Consensus information Everyone in this audience 's reacting in the same way 0 stand-up comedian Cleary, hs routine has worked! consistency is high, and distinctiveness and consensus are also high, one can make an exter nal attribution to the comedian (the cause of Tom’s laughter was the comedian); but where distinctiveness and consensus are low, one can make an internal attribution to Tom’s person- ality (Tom laughed at the comedian because Tom tends to laugh a lot) ‘McArthur (1972) tested Kelley's theory by having participants make internal or external attributions for a range of behaviours, each accompanied by one of the eight possible con- figurations of high or low consistency, distinctiveness and consensus information. Although Consistency Distinetiveness Consensus Attribution Discounting (search fora different cause) Low a External > attribution to the stimulus High + High + High Internal attibution to the person High 4 Low 4 low = > Figure 3.2 Kelley’ attribution theory Kelley covariation model states that people decide what attbutions to make after considering the {@) consistency and (9) distinctiveness of a persons benaviour and (¢) the degree of consensus among, other observers in their reaction to the person's BehaviourEXTENSIONS OF ATTRIBUTION THEORY 89 the theory was generally supported (see review by Kassin, 1979), there was a tendency for people to underuse consensus information. There are also some general issues to consider * Just because people can use pre-packaged consistency, distinctiveness and consensus information to attribute causality (the case in experimental tests of Kelley's model) does not mean that in the normal course of events they do. * There is evidence that people are actually very bad at assessing covariation — they are poor statisticians (Alloy & Tabachnik, 1984). * There is no guarantee that people are using the covariation principle — they may attribute causality to the most salient feature or to whatever causal agent appears to be similar to the effect (Nisbett & Ross, 1980). * IE people do attribute causality on the basis of covariance or correlation, then they cer tainly are naive scientists (Hilton, 1988) - covariation is not causation. Another drawback of the covariation model is that consistency, distinctiveness and con- sensus information require multiple observations. Sometimes we have this information: we may know that Tom does indeed laugh often at almost anything (low distinctiveness), and that others do not find the comedian particularly amusing (low consensus). Ar other times, we may have incomplete information or even no information from multiple observations. How do we now attribute causality? To deal with this, Kelley (1972b) introduced the notion of causal schemata — beliefs or Causa schemata preconceptions, acquired from experience, about how certain kinds of cause interact to pro duce a specific effect. One such schema is that a particular effect requires at least two causes Show (called the ‘multiple necessary cause’ schema): for example, someone with a drunk-driving rere record must have drunk a certain amount of alcohol and have been in control of a vehicle. Although the notion of causal schemata does have some empirical support (Kun & Weiner, 1973) and does help resolve attributional problems raised by the case of a single observation, it is by no means uncritically accepted (Fiedler, 1982) Extensions of attribution theory Explaining our emotions ‘Causal attribution may play a role in how we experience emotions (Schachter, 1964, 1971; for review, see Reisenzein, 1983). Emotions have two distinct components: an undifferentiated state of physiological arousal, and cognitions that label the arousal and determine which emotion is experienced. Usually the arousal and label go hand-in-hand and our thoughts can generate the associated arousal (e.g, identifying a dog as a Rouweiler may produce arousal that is experienced as fear). Sometimes, however, there is initially unexplained arousal that could be experienced as different emotions, depending on what kind of attributions we make for what we are experiencing. This intriguing possibility of ‘emotional lability’ was the focus of a classic study by Schachter and Singer (1962) ~ see Box 3.1 and Figure 3.3. Fora time, the most significant potential of Schachter’s work was the possibility chat it migl be applied in therapy (Valins & Nisbett, 1972). If emotions depend on what cognitive label is, assigned, through causal attribution to undifferentiated arousal, then it might, for example, be possible to transform depression into cheerfulness simply by reattributing arousal, A paradigm. was devised to test this idea — called the misattribution paradigm (Valins, 1966). People who feel anxious and bad about themselves because they attribute arousal internally are encouraged to attribute arousal to external factors. For example, someone who is shy can be encouraged to attribute the arousal associated with meeting new people to ordinary environmental causes rather than to personality deficiencies and thus no longer feel shy. A number of experiments, used this type of intervention with some success (e.g, Olson, 1988; Storms & Nisbett, 1970)90 CHAPTER3 ATTRIBUTION AND SOCIAL EXPLANATION In the late nineteenth century, the famous psychologist William James turned the usual account of how we experi- tence an emotion on its head. As ordinary folk, we might believe that our mental images cause the body to react and thus define our feelings as an emotion. However, James argued that the body first responds automatically to a stimulus, and then we interpret our bodily responses on the basis of what is going on around us: if we see a bear, we run, and a litte later, our pounding heart tells us that we are afraid. One of Stanley Schachter’s experiments dealing with ‘emotional lability’ brought this idea into the laboratory and gave it an attributional flavour (Schachter & Singer, 11962). Male students were given an injection of either adrenalin (the drug epinephrine), or a placebo (salt water) that provided a control condition. Students who had been administered the drug were then allocated to one of three conditions: (1) they were correctly informed that this would cause symptoms of arousal (e.g. rapid breathing, increased heart rate), (2) they were given no explanation ‘or (3) they were misinformed that they might experience a slight headache and some dizziness. All participants then waited in a room with a confederate to complete some paperwork. For half the participants, the confederate behaved euphorically (engaging in silly antics and making paper aeroplanes), and for the other half angrily (ripping up the papers and stomping around). Schachter and Singer predicted that the ‘drug- uninformed’ participants would experience arousal and ‘would search for a cause in their immediate environment (see Figure 3.3). The behaviour of the confederate would act as the salient cue, encouraging participants in the ‘euphoric’ condition to feel euphoric and those in the ‘angry’ condition to feel angry. The emotions of the other two drug groups and the control group would be unaf- fected by the behaviour of the confederate: the control participants had experienced no arousal from the drug, and the correctly informed and misinformed participants already had an explanation for their arousal. The results of the experiment largely supported these predictions However, initial enthusiasm for emotional lability and the clinical application of misat- tribution waned in the light of subsequent criticisms (Buchanan & Seligman, 1995; Forsterling, 1988; Reisenzein, 1983) ‘+ Emotions may be significantly less labile than was originally thought (Maslach, 1979). Environmental cues are not readily accepted as bases for inferring emotions from unex- plained arousal, and because unexplained arousal is intrinsically unpleasant, people have a propensity to assign it a negative label ‘© The misattribution effect is unreliable, short-lived and largely restricted to laboratory studies (Parkinson, 1985). It is not clear that itis mediated by an attribution process, and in any case it is also restricted to a limited rage of emotion-inducing stimuli. Immediate Bodily [Appraisal of Subjective stimulus response context emotion ‘other person ine _ = ‘Arousal ‘Adrenalin injected > (nerensein ‘ ‘heart rate) os > Euphoria isfunny Figure 3.3 Attributing a likely cause to an experimentally induced emotionEXTENSIONS OF ATTRIBUTION THEORY 91 The more general idea that cognition, particularly cognitive appraisals of the surround- ing situation, plays an important role in generation and experience of emotion has, however, fed into the contemporary revival of research on affect and emotion (e.g, Blascovich, 2008; Forgas, 2006; Forgas & Smith, 2007; Haddock & Zana, 1999; Keltner & Lerner, 2010; see Chapter 2). Indeed, attribution theory was the conceptual springboard for the later explora tion of the concept of appraisal (e.g, Lazarus, 1991). Attributions for our own behaviour ‘One significant implication of treating emotion as cognitively labelled arousal is the possi bility that people make more general attributions for their own behaviour. This idea has been elaborated by Daryl Bem (1967, 1972) in his self-perception theory. (Because this is an account of how people construct their self-concept, we deseribe it in Chapter 4 which explores the nature of self and identity) Task performance attributions Another extension of attribution theory focuses on the causes and consequences of the attri bution people make for how well they and others perform on a task — for example, success or failure in an examination (Weiner, 1979, 1985, 1986). In making an achievement attribu- tion, we consider three performance dimensions 1. Locus ~is the performance caused by the actor (internal) or by the situation (external)? 2 Stability is the internal or external cause a stable or unstable one? 3 Controllability —to what extent is future task performance under the actor’s control? These produce eight different types of explanation for task performance (see Figure 3.4) For example, failure in an examination might be attributed to ‘unusual hindrance from others’ {the top right-hand box in Figure 3.4 ifthe student was intelligent (therefore, failure is external) Internal External Stable Unstable Stable Unstable “enwewie ea te = = rene st to set a Figure 3.4 Achievement attributions as a function of locus, stability and controllability How we acribute someone's task achievement dezends on * Locus ~ isthe performance caused by the actor (internal) or the situation (external? * Stability ~ is the inceral or exernal cause a stable or unstable one? © Contaliability ~ to what extent is future task performance under the actor's contol? Selt-perception theory Bemisia that we gar knowledge of ouseives only oy making seatibutons for eramale, we ler our fw attudes Fam our on behaviour92. CHAPTER3 ATTRIBUTION AND SOCIAL EXPLANATION. [Atributional style Achievement attribution ‘will these Reality TV incesnal (personal) factors or to excernal {situational factors? and was disturbed by a nearby student sneezing from hay fever (anstable and controllable, because in future examinations the sneezing student might not be present or have taken an anti-histamine, and/or one could choose to sit in a place away from the sneezing student) cording to Weiner, people first determine whether someone has succeeded or failed and accordingly experienced positive or negative emotion. They then make a causal attribution for the performance, which produces more specific emotions (e.g, pride for doing well due to ability) and expectations that influence future performance This idea is relatively well supported by experiments where participants are provided with performance outcomes and locus, stability and controllability information, often under role-playing conditions (e.g. De Jong, Koomen, & Mellenbergh, 1988; Frieze & Weiner, 1971). However, critics have suggested that controllability may be less important than was first thought and have wondered to what extent people outside controlled laboratory condi- tions really analyse achievement in this way. Subsequently, Weiner (1995) has placed an emphasis on judgements of responsibility. On the basis of causal attributions, people make tudgements of responsibility, and these latter judgements, not the causal attributions them- selves, influence affective experience and behavioural reactions. Applications of attribution theory Application of the idea that people need to discover the cause of their own and others behaviour in order to plan their own actions has had a significant impact on social psychol. ogy. We have already seen two examples — achievement attributions and the reattribution of, arousal as a therapeutic technique. Here, we explore two further applications: attributional styles and interpersonal relationships. Individual differences and attributional styles Research suggests that people differ in the sorts of attributions they make; they have differ ent attributional styles. This is because they differ in the amount of control they feel they have over the reinforcements and punishments they receive (Rotter, 1966). Internals believe they have significant personal control over their destiny ~ things happen because they make le control over them happen. Externals are more fatalistic ~ they believe that they have lit what happens to them; things simply occur by chance, luck or the actions of powerful[APPLICATIONS OF ATTRIBUTION THEORY 93 external agents, To measure people’s locus of control, Rotter devised a twenty-nine-item scale, This scale has been used to relate locus of control to a range of behaviours, including political beliefs, achievement behaviour and reactions to illness. One problem with the scale is that it may measure not a unitary construct (i.e. a single personality dimension) but, rather, a number of relatively independent beliefs to do with control (Collins, 1974) “The notion of individual differences in attributional style, a tendency for individuals to make particular kinds of causal inference rather than others, over time and across different situations, has sponsored the development of a number of questionnaires to measure attribu- tional style (Metalsky & Abramson, 1981). Of these, the attributional style questionnaire or ASQ (Peterson et al., 1982; Seligman, Abramson, Semmel, &¢ Von Baeyer, 1979) is perhaps the most widely known, It measures the sorts of explanation that people give for aversive (ise. unpleasant) events on three dimensions: internaliexternal, stable/unstable and global/specifc. The global/specific dimension refers to how wide or narrow a range of effects a cause has ~ “the economy” is a global explanation for someone being made redundant, whereas the clos. ing of a specific company is a specific explanation. People who view aversive events as being caused by internal, stable, global factors have a ‘depressive attributional style’ (ie. the glass is, half empty), which may promote helplessness and depression and may have adverse health. consequences (Abramson, Seligman, & Teasdale, 1978; Crocker, Alloy, & Kayne, 1988) Another, slightly different scale, called the attributional complexity scale (ACS), has been devised by Fletcher et al. (1986) to measure individual differences in the complexity of the attributions that people make for events. ‘The idea that attributional style is personality trait is not without problems: for instance, the ASQ and the ACS provide only limited evidence of cross-situational individual consist- ency in causal attribution (e.g. Cutrona, Russel, & Jones, 1985). Also not without problems is the link between attributional style, learnt helplessness and clinical depression. Although more than 100 studies involving about 15,000 participants confirm an average correlation of 0.30 between attributional style and depression (Sweeney, Anderson, & Bailey, 1986), this, does not establish causation —it is a correlation where one factor explains 9 per cent of vari- ance in the other. ‘More useful are studies that show that attributional style measured at one time predicts depressive symptoms at a later date (Nolen-Hoeksma, Girgus, &¢ Seligman, 1992), but again, causality is not established. Causality is difficult to establish because it is of course unethical to induce clinical depression in experimental settings. We are largely left with experimental evidence from studies of transitory mood, which is a rather pale analogue of depression. Is it justified to generalise from feelings about doing well or poorly on a trivial laboratory task to full-blown clinical depression? Interpersonal relationships Attributions play an important role in interpersonal relationships (see Chapter 14); particu larly close relationships (e.g. friendship and marriage) where people communicate attribu- tions, for example to explain, justify or excuse behaviour and to attribute blame and instil guilt (Hilton, 1990) Interpersonal relationships typically go through three basic phases: formation, maintenance and dissolution (Harvey, 1987; see also Moreland and Levine's (1982; Levine & Moreland, 1994) model of group socialisation in Chapter 8). During the formation stage, attributions reduce ambiguity and facilitate communication and an understanding of the relationship (Fincham, 1985). In the maintenance phase, the need to make attributions wanes because sta- ble personalities and relationships have been established. The dissolution phase is character- ised by an increase in attributions in order to regain an understanding of the relationship. A not-uncommon feature of interpersonal relationships is attributional conflict (Horai, 197), where partners proffer divergent causal interpretations of behaviour and disagree94 CHAPTERS ATTRIBUTION AND SOCIAL EXPLANATION Attributing blame Couples sometimes cannot agree on what is cause and what is effect. For example, does nagging cause withérawal or withdrawal cause nageng? over what attributions to adopt. Often partners cannot even agree on a cause-effect sequence, one exclaiming, ‘I withdraw because you nag’, the other, ‘I nag because you with- draw’. Research mainly on heterosexual couples has shown that attributional conflict is, strongly associated with relationship dissatisfaction (Kelley, 1979; Orvis, Kelley, & Butler, 1976; Sillars, 1981). ‘Most research has focused on the role of attributions in heterosexual marital satisfaction (e.g, Fincham & Bradbury, 1991; Fletcher & Thomas, 2000; Noller & Ruzzene, 1991), wi the aim of distinguishing between distressed and non-distressed spouses in order to provide therapy for dysfunctional marital relationships. Correlational studies (c.g, Fincham & O'Leary, 1983; Holtzworth-Muntoe, & Jacobson, 1985) reveal that happily married (or non-distressed) spouses tend to credit their partners for positive behaviour by citing internal, stable, global and controllable factors to explain them, Negative behaviour is explained away by ascribing it to causes viewed as external, unstable, specific and uncontrollable. Distressed couples behave in exactly the opposite way While women fairly regularly think in causal terms about the relationship, men do so only when the relationship becomes dysfunctional. In this respect, and contrary to popular opin- on, men may be the more diagnostic barometers of marital dysfunction ~ when men start analysing the relationship, alarm bells should ring? Do attributional dynamics produce dysfunctional marital relationships, or do dysfunc- tional relationships distort the attributional dynamic? This key causal question has been addressed by Fincham and Bradbury (1987; see overview by Hewstone, 1989}, who measured responsibility attributions, causal attributions and marital satisfaction in 39 married cou- ples on two occasions 10-12 months apart. Attributions made on the first occasion were found reliably to predict marital satisfaction 10-12 months later, but only for wives. Another longitudinal study (although over only a two-month period) confirmed that attributions do have a causal impact on subsequent relationship satisfaction (Fletcher, Fincham, Cramer, & Heron, 1987). Subsequent, more extensive and better-controlled longi- tudinal studies have replicated these findings for both husbands and wives (Fincham & Bradbury, 1993; Senchak & Leonard, 1993) Attributional biases ‘The attribution process is clearly subject to bias: for example, it can be biased by personality, biased by interpersonal dynamics or biased to meet communication needs. We do not approach the task of attributing causes for behaviour in an entirely dispassionate,ATTRIBUTIONALBIASES 95 disinterested and objective manner, and the cognitive mechanisms that are responsible for attribution may themselves be subject to imperfections that make them suboptimal Asevidence of attributional biases and ‘errors’ accumulated, there was a shift of perspec- tive. Instead of viewing people as naive scientists or even statisticians (in which case biases were largely considered a theoretical nuisance), we now think of people as cognitive misers or motivated tacticians (Moskowitz, 2005; Fiske & Taylor, 2013; see Chapter 2). People use cognitive short-cuts (called heuristics) to make attributions that, although not always accu- rate oF correct, are quite satisfactory and adaptive. Sometimes the choice of short-cut and choice of attribution can also be influenced by personal motives, Biases are entirely adaptive characteristics of ordinary, everyday social perception (Fiske & Taylor, 2013; Nisbett & Ross, 1980; Ross, 1977). In this section, we discuss some of the most important attributional biases. Correspondence bias and the fundamental attribution error ‘One of the best-known attribution biases is correspondence bias — a tendency for people to over-attribute behaviour to stable underlying personality dispositions (Gilbert & Malone, 1995) (see Box 3.2). This bias was originally called the fundamental attribution error. Although the correspondence bias and fundamental attribution errors are not identical (Gawronski, 2004), the terms are often used interchangeably ~ the change in the preferred Jabel mainly reflects evidence that this bias or error may not be quite as ‘fundamental’ as originally thought (sce ‘Cultural and developmental factors’) ‘The fundamental attribution error, originally identified by Ross (1977), is a tendency for people to make dispositional attributions for others’ behaviour, even when there are clear external/environmental causes, For example, in the Jones and Harris (1967) study mentioned carlier, American participants read speeches about Cuba’s President Fidel Castro ostensibly written by fellow students. The speeches were either pro-Castro or anti-Castro, and the weit- crs had ostensibly cither freely chosen to write the specch or been instructed to do so. Where there was a choice, participants not surprisingly reasoned that those who had written a pro Castro speech were in favour of Castro, and those who had written an anti-Castro speech were against Castro — an internal, dispositional ateribution was made (see Figure 3.5). However, a dispositional attribution was also made even when the speech-writers had been instructed to write the speech. Although there was overwhelming evidence for an exclu- sively external cause, participants largely disregarded this and still preferred a dispositional explanation — the fundamental attribution error. (Bearing these points in mind, how would Cognitive miser mode ot sacs cogntion that characterises people using the least complex and emancing cognitions that axe able to produce eenenaly adaptive Behaviours Motivate tactician Amode of socal cognsion that characterises people as having multiple cognitive strates avalable, which ‘hay choose among onthe bass of personal goals motives ane needs. Correspondence bias A general ateibuon bias in hich people have an Inflatedtendency to see behaviour as eetng (corresponding ta) stable underlying personality atibutes. Fundamental attribution Bian atauing nother behawiour more to internal thant st. asonal causes People's inherent tendency to fall prey to the corre- spondence bias can be exploited by the politcal process. In the 2016 US presidential election, the Republican Party spun information about the past behaviour of the Democratic contender, Hillary Clinton, to paint @ picture of heras an untrustworthy and unlikable person - Donald ‘Trump, the Republican contender, repeatedly used the term ‘crooked Hillary’. The Democratic Party, in turn, drew attention to Trump's behaviour (his tweets and ‘campaign speeches), to paint a picture of him as an Unstable, thin-skinned narcissist dangerously unsuited to ‘the presidency. In both cases the partisan electorate seemed more comfortable focusing on the flawed personality of the opposing presidential contender than on the more com- plex policy landscape of the party the contender repre- sented. When an election ‘gets personat’ by focusing on and overinflating or falsely creating an opponent's per- sonal failings, it plays right into the hands of the corre- spondence bias and ultimate attribution error,96 CHAPTER3 ATTRIBUTION AND SOCIAL EXPLANATION Pro-Castro 75 3% Figure 35 The fundamental attribution error: = [Attributing speech writes’attitudes on the basis g of their freedom of choice in writing the speech Sos ‘Students who freely chose to write a pro- or an ant-Castro speech were attributed with a pro- or anvisCasteo attitude respectively. ‘though less trong, this same tendency ta attribute 1 Pro-casto speech BE Aniicasto speech the speech to an underlying dupostion (he Aniscasro 0 fundamental atibuton er) prevaled when the Choice No Choice writers had no choice and were simply instruct to Degree ofchoice weite the so Source: Based on data om Jone and Mars (1967) Essential Perasveendency 20 consider behavour to reer underjing and Immutable often innte, properties of people or the ‘groups they belong to you account for the different views held by Helen and Lewis? See the first ‘What do you think’ question.) The fundamental attribution error, or correspondence bias, has been demonstrated repeatedly both inside and outside the social psychology laboratory (Gawronski, 2004; Gilbert, 1998; Jones, 1979, 1990; Nisbett & Ross, 1980). Correspondence bias may also be responsible for a number of more general explanatory tendencies: for example, people's tendency to attribute road accidents to the driver rather than to the vehicle or the road con- ditions (Barjonet, 1980}; and some people’s tendency to attribute poverty and unemploy- ‘ment to the person rather than to social conditions (sec the discussion of the key term ‘Belief in a just world? later in this chapter) Pettigrew (1979) has suggested that the fundamental attribution error may emerge in a slightly different form in intergroup contexts where groups are making attributions about ingroup and outgroup behaviour ~ he calls this the ultimate attribution error (sce the Intergroup Attribution’ section later in this chapter). Correspondence bias and the funda- mental attribution error are closely related to two other biases: the outcome bias (c.g Allison, Mackie, & Messick, 1996), where people assume that a petson behaving in some particular way intended all the outcomes of that behaviour; and essentialism (Haslam, Rothschild, & Ernst, 1998; Medin & Ortony, 1989), where behaviour is considered to reflect underlying and immutable, often innate, properties of people or the groups they belong to. Essentialism can be particularly damaging when it causes people to attribute stereotypi- cally negative attributes of outgroups to essential and immutable personality attributes of members of that group (e.g. Bain, Kashima, & Haslam, 2006; Haslam, Bastian, Bain, 8 Kashima, 2006; Haslam, Bastian, & Bissett, 2004). There is evidence that groups can use essentialism strategically to discriminate against outgroups (Morton, Hornsey, & Postmes, 2009). For example, the stereotype of an outgroup as being laid-back, liberal and poorly educated becomes more pernicious if these attributes are considered immutable, perhaps genetically induced, properties of the group's members ~ the people themselves are consid- ered to have personalities that are immutably lazy, immoral and stupid. Different explanations of the correspondence bias have been proposed. They include: 1. Focus of attention. The actor's behaviour attracts more attention than the backgrounds itis, disproportionately salient in cognition, stands out as the figure against the situational back- ground and is therefore over-represented causally (Taylor & Fiske, 1978). Thus, the actorATTRIBUTIONAL BIASES and the actor's behaviour form what Heider (1958) called a ‘causal unit’. This explanation makes quite a lot of sense. Procedures designed to focus attention away from the actor and on to the situation inerease the tendency to make a situational rather than dispositional attribution (e.g, Rholes & Pryor, 1982). When people really want to find out about a situ- ation from a person's behaviour, they focus on the situation and ate less likely to leap to a dispositional attribution ~ the correspondence bias is muted or reversed (e.g. Krull, 1993) 2 Differential forgetting. Attribution requires the representation of causal information in memory. There is evidence that people tend to forget situational causes more readily than dispositional causes, thus producing a dispositional shift over time (e.g. Moore, Sherrod, Liu, & Underwood, 1979; Peterson, 1980). Other studies show the opposite effect (e.g, ‘Miller & Porter, 1980), and Funder (1982) has argued that the direction of shift depends on the focus of information processing and occurs immediately after the behaviour being attributed, 3. Linguistic facilitation. One rather interesting observation by Nisbett and Ross (1980) is that the construction of the English language makes it relatively easy to describe an action and the actor in the same terms, but more difficult to describe the situation in the same way. For example, we can talk about a kind or honest person, and a kind or honest action, but not a kind or honest situation. The English language may facilitate disposi- tional explanations (Brown & Fish, 1983; Semin & Fiedler, 1991) Cultural and developmental factors The correspondence bias was originally called the fundamental attribution error because it was considered an automatic and universal outcome of perceptual experience and cognitive activity (e.g. McArthur & Baron, 1983). However, there is evidence that both developmental, factors and culture may affect the correspondence bias. For example, in Western cultures, young children explain action in concrete situational terms and learn to make dispositional attributions only in late childhood (Kassin & Pryor, 1985; White, 1988). Furthermore, this developmental sequence itself may not be universal. Hindu Indian children do not drift towards dispositional explanations at all, but rather towards increasingly situational expla- nations (Miller, 1984). We return to this point later when we discuss particular cultural and developmental differences in how children make attributions (see Figure 3.7). These differences quite probably reflect different cultural norms for social explanation, or more basic differences between Western and non-Western conceptions of self ~ the autono- mous and independent Western self and the interdependent non-Western self (Chiu & Hong, 2007; see Chapters 4 and 16), ‘The correspondence bias is a relatively ubiquitous and socially valued feature of Western cultures (Beauvois & Dubois, 1988; Jellison & Green, 1981), but, although present, itis less dominant in non-Western cultures (Fletcher & Ward, 1988; Morris & Peng, 1994) ‘As noted carlicr, the fundamental attribution error is not as fundamental as was origi- nally thought. In many ways, it may be a normative way of thinking (see discussion of norms in Chapters 7 and 8). Thisis one reason why Gilbert and Malone (1995) recommend that the term ‘correspondence bias’ be used in preference to the term ‘fundamental attribution error’. Indeed, according to Gawronski (2004), the two constructs are subtly different: technically, he argues, the fundamental attribution error is the tendency to underestimate the impact of situational factors; and the correspondence bias is the tendency to draw correspondent dis- positional inferences from behaviour that is constrained by the situation. The actor-observer effect Imagine the last time a shop assistant was rude to you. You probably thought, ‘What a rude person!’ although perhaps put less politely ~ in other words, you made an internal attribu- tion to the shop assistant’s enduring personality. In contrast, how did you explain the last 798 CHAPTER3 ATTRIBUTION AND SOCIAL EXPLANATION. ‘Actor-observer effect Tencencyt 3 ann behavus externally and others behaviours time you snapped at someone? Probably not in terms of your personality; more likely in of external factors such as time pressure or stress. The actor-observer effect er effect) is really an extension of the correspondence bias. It refers to the tendency (or the for people to attribute others’ behaviour internally to dispositional factors and their own behaviour externally to environmental factors (Jones & Nisbett, 1972) Research has provided substantial evidence for this effect (Watson, 1982), and some extensions and qualifications, For example, not only do we attribute others’ behaviour more dispositionally than our own, but we also consider their behaviour to be more stable and predictable than our own (Baxter & Goldberg, 1988). The valence of the behaviour also matters. People make more dispositional attributions for socially desirable than socially undesirable behaviour, irrespective of who the actor is (e.g, Taylor & Koivumaki, 1976), and actors are more dispositional in attributing positive behaviour and more situational in attrib- uting negative behaviour than are observers (e.g. Chen, Yates, & McGinnies, 1988) The actor-observer effect can be inverted if someone knows their behaviour is disposi tionally caused. For example, you may ‘adopt’ an injured hedgehog knowing that you are a sucker for injured animals and you have often done this sort of thing in the past (Monson & Hesley, 1982). Finally, the actor-observer effect can be erased or reversed if the actor is encouraged to take the role of the observer regarding the behaviour to be attributed, and the observer the role of the actor. Now the actor becomes more dispositional and the observer more situational (e.g. Frank & Gilovich, 1989) There are two main explanations for the actor—observer effect: 1 Perceptual focus. This explanation is almost identical to the ‘focus of attention’ expla- nation for the correspondence bias described earlier in this chapter. For the observer, the actor and the actor’s behaviour are figural against the background of the situation However, actors cannot ‘see’ themselves behaving, so the background situation assumes the role of figure against the background of self. The actor and the observer quite liter- ally have different perspectives on the behaviour and therefore explain it in different ways (Storms, 1973). Perceptual salience does indeed seem to play an important role in causal explanation. For example, McArthur and Post (1977) found that observers made more dispositional attributions for an actor’s behaviour when the actor was strongly illumi- nated than when dimly illuminated 2 Informational differences. Another reason why actors tend to make external attributions and observers internal ones is that actors have a wealth of information to draw on about how they have behaved in other circumstances. They may actually know that they behave differently in different contexts and thus quite accurately consider their behaviour to be under situational control. Observers are not privy to this autobiographical information, ‘They see the actor behaving in a certain way in one context, or a limited range of contexts, and have no information about how the actor behaves in other contexts. Its therefore not an unreasonable assumption to make a dispositional attribution. This explanation, first suggested by Jones and Nisbett (1972), does have some empirical support (Eisen, 1979; White & Younger, 1988). The false consensus effect Kelley (1972a} identified consensus information as being one of the three types of informa: tion that people used to make attributions about others’ behaviour (see earlier in this chap- ter). One of the first cracks in the naive scientist model of attribution was McArthur’s, (1972) discovery that attributors in fact underused or even ignored consensus information (Kassin, 1979). Subsequently, it became apparent that people do not ignore consensus information but rather provide their own consensus information. People see their own behaviour as typi- cal and assume that, under similar circumstances, others would behave in the same way:ATTRIBUTIONALBIASES 99 The false consensus effect This mid-winter arctic dipper discovers an altributional bias, Who else would swim here before breakfast? Ross, Greene and House (1977] first demonstrated this false consensus effect. They asked false consensus eect students if they would agree to walk around campus for 30 minutes wearing a sandwich _Se2rgou oun beravou board carrying the slogan ‘Eat at Joe’s’. Those who agreed estimated that 62 per cent of | 267émaewypesl neni their peers would also have agreed, while those who refused estimated that 67 per cent of their peers would also have refused. ‘Well over 100 studies testify to the robust nature of the false consensus effect (Marks & Miller, 1987; Mullen, Atkins, Champion, Edwards, Hardy, Story, 8 Vanderklok, 1985; Wetzel & Walton, 1985). The effect exists for a number of reasons: © We usually seek out similar others and so should not be surprised to find that other peo- ple are similar to us. * Our own opinions are so salient to us, at the forefront of our consciousness, that they eclipse the possibility of alternative opinions. * We are motivated to ground our opinions and actions in perceived consensus in order to validate them and build a stable world for ourselves. The false consensus effect is stronger for important beliefs, ones that that we care a geeat deal about (e.g. Granberg, 1987), and for beliefs about which we are very certain (e.g. Marks & Miller, 1985). In addition, external threat, positive qualities, the perceived similarity of others and minority group status also inflate perceptions of consensus (c.g. Sanders & Mullen, 1983; Sherman, Presson, & Chassin, 1984; Van dee Pligt, 1984) Self-serving biases In keeping with the motivated tactician model of social cognition (Fiske & Taylor, 1991) discussed earlier in this chapter (also see Chapter 2), attribution is influenced by our desire for a favourable image of ourselves (see Chapter 4). We make attributions that satisfy self-serving biases. Overall, we take ctedit for our positive behaviours and successes as SS"sewingbiaies reflecting who we are and our intention and effort to do positive things (the self-enhancing yawsr sy onnaree sel bias}, At the same time, we explain away our negative behaviours and failures as being due _ escem orthe seleconcep.100 CHAPTER3 ATTRIBUTION AND SOCIAL EXPLANATION sethandicapping Pally mating adva enteral aebations anucpated fale or fonhcoming event Huson of contrt Belt that we have Belief in ajust world Bole hat the wor ju and precicable pace wi 000 things happen so ‘00d people and esto" to coercion, normative constraints and other external situational factors that do not reflect who we ‘really’ are (the self-protecting bias). This is a robust effect that holds across many cultures (Fletcher & Ward, 1988) Sclf-serving biases are clearly ego-serving (Snyder, Stephan, & Rosenfield, 1978). However, Miller and Ross (1975) suggest there is also a cognitive component, particularly for the self- enhancing aspect. People generally expect to succeed and therefore accept responsibility for suceess. If they try hard to succeed, they associate success with their own effort, and they generally exaggerate the amount of control they have over successful performances. Together, these cognitive factors might encourage internal attribution of suecess, Overall, it is most likely both cognitive and motivational factors have a role (Anderson & Slusher, 1986; Tetlock & Levi, 1982} and they are difficult to disentangle from one another (Tetlock & ‘Manstead, 1985; Zuckerman, 1979) Self-enhancing biases are more common than self-protecting biases (Miller & Ross, 1975} ~ partly because people with low self-esteem tend not to protect themselves by attributing their failures externally; rather, they attribute them internally (Campbell & Fairey, 1985) Both of these forms of bias can be muted by a desire not to be seen to be boasting over our successes and lying about our failures (e.g. Schlenker, Weingold, 8& Hallam, 1990) — but they are not totally extinguished (Riess, Rosenfield, Melburg, & Tedeschi, 1981). One self-serv ing bias which most of us have exploited from time to time is self-handicapping, a term described by Jones and Berglas: The self-handicapper, we are suggesting, reaches out for impediments, exaggerates handi- caps, embraces any factor reducing personal responsibility for mediocrity and enhancing personal responsibility for success Jones and Berglas (1978, p. 202) People self-handicap in this way when they anticipate failure, whether in their job perfor: mance, in sport, or even in therapeutic settings when being ‘sick’ allows one to drop out of life. What a person often will do is intentionally and publicly make external attributions for a poor showing even before it happens. Check the experiment about choosing between drugs n Box 3.3 and Figure 3.6. Another instance of self-serving attribution surfaces when attribution of responsibility (Weiner, 1995) is influenced by an outcome bias (Allison, Mackie, & Messick, 1996). People tend to attribute greater responsibility to someone who is involved in an accident with large rather than small consequences (Burges, 1981; Walster, 1966). For example, we would attrib- ute greater responsibility to the captain of a super-tanker that spills millions of litres of oil than to the captain of a charming little fishing boat that spills only a few litres, although the degree of responsibility may actually be the same. ‘This effect quite probably reflects the tendency for people to cling to an illusion of control (Langer, 1975) by believing in a just world (Furnham, 2003; Lerner, 1977). People like to believe that bad things happen to ‘bad people” and good things to ‘good people’ (i.e. people get what they deserve), and that people have control over and responsibility for their out- comes. This attributional pattern makes the world seem a controllable and secure place in which we can determine our own destiny. Belief in a just world can result in a pattern of attribution where victims are deemed responsible for their misfortune — poverty, oppression, tragedy and injustice all happen. because victims deserve it, Examples of the just world hypothesis in action are such views as the unemployed are responsible for being out of work, and rape victims are responsible for the violence against them. Another example is the belief, still held by some people, that the 6 million Jewish victims of the Holocaust were responsible for their own fate ~ that they deserved it (Davidowicz, 1975), Refer back to the second ‘What do you think?” question. Just world beliefs are also an important component of many religious ideologies (Hogg, Adelman, & Blagg, 2010}Box 3.3 Your life Self-handicapping: Explaining away your failure Imagine that you are waiting to take an examination in a subject you find difficult and that you fully anticipate fail- ing. You might well make sure that as many people as pos- sible know that you have done no revision, are not really interested in the subject and have a mind-numbing hang- ‘over to boot. Your subsequent failure is thus externally attributed without it seeming that you are making excuses to explain away your failure. To investigate this idea, Berglas and Jones (1978) had introductory psychology students try to solve some prob- lems where the problems were either solvable or not solv- able. They were told that they had done very well, and 100 1 Actavil- improves performance IH Pandocrin- hinders performance Percentage of subjects choosing each drug bs 8 Not solvable Type of puzzle worked on Solvable INTERGROUP ATTRIBUTION 101 before continuing with a second problem-solving task, ‘they were given the choice of taking either a drug called ‘Actavil, which would ostensibly improve intellectual func- tioning and performance, or 'Pandocrin, which would have the opposite effect. As predicted, those students who had succeeded on the solvable puzzles felt confident about their ability and so chose Actavil in order to improve further (see Figure 3.6). Those who had succeeded on the not-solvable puzzles attributed their performance exter- nally to luck and chose Pandocrin in order to be able to explain away more easily the anticipated failure on the second task Figure 3.6 Self-handicapping: Choosing a drug depends on a puzzle’ solvability ‘Students wno had done well on a solvable puzzle could attribute their performance internally (eg, t0 abil) Anticipating an equally good performance on a second similar task, they cnose a performance-enhancing drug, ‘Actavi rather than a performance-impairing drug, Pandocrin, ‘Students whe had done well on a not-solvable puzzle could only attribute ther perfarmance extemaly (e.g t0 luck}: with the prospect of an equivalent performance on the second task they chose the per‘ormance-impairing drug, asthe sel-handicapping option, Source: ses on data from Berga ana Janes (1578) Belief in a just world may also be responsible for sel-blame. Victims of traumatic events such as incest, debilitating illness, rape and other forms of violence can experience a strong sense that the world is no longer stable, meaningful, controllable or just. One way to reinstate an illu- sion of control is, ironically, to take some responsibility for the event (Miller & Porter, 1983}. Intergroup attribution Attribution theories are concerned mainly with how people make dispositional or situational attributions for their own and others’ behaviour and the sorts of bias that distort this process. ‘The perspective is tied to interpersonal relations: people as unique individuals make attributions102 Intergroup atribution Process of asin ‘CHAPTER 3 Ethnocentrism falas [ATTRIBUTION AND SOCIAL EXPLANATION Intergroup attribution Black Lives Matter isan international movernent of political activists ‘whose belief system incoraorates social charge {see Chapter 11) Treir ouigroup is nuge and ‘governments, po security forces their own behaviour or the behaviour of other unique individuals. However, there is another autributional context ~intergroup relations — where individuals as group members make attribu- tions for the behaviour of themselves as group members and others as either ingroup or out- group members (Deschamps, 1983; Hewstone, 1989; Hewstone & Jaspars, 1982, 1984) Examples of intergroup attribution abound. One example is the attribution of national economic and social malaise to immigrant minorities (e.g. Middle Eastern and North African refugees in Europe, Eastern Europeans in the United Kingdom and Mexicans in the United States). Another is the explanation of behaviour in terms of stereotypical properties of a person’s group membership — for example, attributions for performance that are con. sistent with gender or racial stereotypes (Inzlicht & Schmader, 2011} Intergroup attributions serve two functions, the first relating to ingroup bias and the sec- ond to self-esteem. Extending our discussion of self-serving attributional biases to inter- group relations, ethnocentrism can be viewed as an ingroup-serving bias. Socially desirable {positive} behaviour by ingroup members and socially undesirable (negative) behaviour by spessofou'chn EMP Gutgroup members ate internally attributed to dispositions, and negative ingroup and posi- . ‘ tive outgroup behaviour are externally attributed to situational factors Jaspars, 1982; Hewstone, 1989, 1990). This tendency is more prevalent in Western than in non-Western cultures (Fletcher & Ward, 1988) and is common in team sports contexts, where the success of one’s own team is attributed to internal stable abilities rather than effort, luck or task difficulty — we ate skilful, they were lucky. This group-enhancing bias is stronger and more consistent than the corresponding group-protective bias (Mullen & Riordan, 1988; Miller & Ross, 1975). Ubimateateibution ewer Pettigrew (1979) has described a related bias, called the ultimate attribution error. This is Terseoyaioustna an extension of Ross's (1977) fundamental attribution ertor that focuses on attributions for behaviour inernly and to atibute ood outgroup xd ba outgroup behaviour, Pettigrew argued that negative outgroup behaviour is dispositionally attributed, whereas positive outgroup behaviour is externally attributed or explained away ngroupbenviow so that we preserve our unfavourable outgroup image. The ultimate attribution error refers excel to attributions made for outgroup behaviour only; whereas broader intergroup perspectives focus on ingroup attributions as well Taylor and Jaggi (1974) conducted an early study of intergroup attributions in southern India against a background of intergroup conflict between Hindus and Muslims. Hindu participants read vignettes describing Hindus or Muslims acting in a socially desirable way (e.g. offering shelter from the rain) or socially undesirable way (e.g, refusing shelter) towards them and then chose one of a number of explanations for the behaviour. The results were as, predicted. Hindu participants made more internal attributions for socially desirable thanINTERGROUP ATTRIBUTION 103, socially undesirable acts by Hindus (ingroup). This difference disappeared when Hindus made attributions for Muslims (outgroup) Hewstone and Ward (1985) conducted a more complete and systematic follow-up, with Malays and Chinese in Malaysia and Singapore. Participants made internal or external attributions for desirable or undesirable behaviour described in vignettes as being per~ formed by Malays or by Chinese. In Malaysia, Malays showed a clear ethnocentric attribu- tion bias ~ they attributed a positive act by a Malay more to internal factors than a similar act by a Chinese, and a negative act by a Malay less to internal factors than a similar act by a Chinese (see Figure 3.7). The ingroup enhancement effect was much stronger than the outgroup derogation effect. The Chinese participants showed no ethnocentric bias instead, they showed a tendency to make similar attributions to those made by Malays. In Singapore, the only significant effect was that Malays made internal attributions for positive acts by Malays, Hewstone and Ward explain these findings in terms of the nature of intergroup relations in Malaysia and Singapore, In Malaysia, Malays are the clear majority group and Chinese an ethnic minority. Furthermore, relations between the two groups were tense and relatively conflictual at the time, with Malaysia pursuing a policy of ethnic assimilation. Both Malays and Chinese generally shared an unfavourable stereotype of Chinese and a favourable ste- reotype of Malays. In contrast, Singapore has been ethnically more tolerant. ‘The Chinese are in the majority, and ethnic stereotypes are markedly less pronounced. 100 1 Negative act I Positive act In Mali InSienpore pep 3 # oa z z 5 40 Fon 99 Target: Chinese Malay Chinese Malay Chinese Malay Participant: CHINESE MALAY CHINESE Ethnicity of targets and paticipants Figure 3.7 Internal attribution of positive and negative acts by Malays or Chinese as a function of attributor ethnicity, ‘Malays showed an ethnocentric attributional bias in which a postive act was more interally attributed to a ‘Malay than a Chinese, and a negative act less internally attributed to a Malay than a Chinese: the effect was more pronounced in Malays, where Malays are the dominant group and Chinese the ethnic minor ty, than in Singapore. Chinese did not show an ethnocentric attribution bias “Source Bed on data tom Henstone ane Ward 1985) Stereotype Weel sharee and simpifed evaluative image (of socal group ands members104 CHAPTER3 ATTRIBUTION AND SOCIAL EXPLANATION vecnansms and language used to expiaina Social identity theory Theory of membership and intergroup reatons based on sa separation, oc comparison and the constuction of shared ngroup-derining properties ‘The takeaway message from this analysis is that ethnocentric attribution is not a universal that reflects asocial cognition; rather, it depends on intergroup dynamics in a soci- historical context. The sorts of attribution that group members make about ingroup and outgroup behaviour are influenced by the nature of the relations between the groups. ‘This is consistent with Hewstone's (1989) argument that a fuller analysis of attribution, more accurately described as social explanation, requires a careful articulation (i.e. theoreti- cal integration or connection) of different levels of explanation (see Doise, 1986; see also Chapter 1), In other words, we need to know how individual cognitive processes, interper- sonal interactions, group membership dynamics and intergroup relations all affect, are affected by and are interrelated with one another. Further evidence for ethnocentric intergroup attributions comes from studies of inter- racial attitudes in educational settings in the United States (Duncan, 1976; Stephan, 1977)s, from studies of inter-ethnic relations between Israelis and Arabs (Rosenberg & Wolfsfeld, 1977} and between Hindus and Muslims in Bangladesh (Islam & Hewstone, 1993); and from studies of race, sex and social class-based attributions for success and failure (Deaux. & Emswiller, 1974; Feather & Simon, 1975; Greenberg & Rosenfield, 1979; Hewstone, Jaspars, & Lalljec, 1982} ‘More recently, Mackie and Ahn (1998) found that the outcome bias, the assumption that the outcomes of behaviour were intended by the person who chose the behaviour, is affected by whether the actor is a member of your group oF not, and whether the outcome was desir able or not. Mackie and Ahn found that there was an outcome bias in the case of an ingroup member and a desirable outcome but not when the outcome was undesirable. At least two processes may be responsible for ethnocentric intergroup attributions: 1A cognitive process: Social categorization generates category-congruent expectations in the form of expectancies (Deaux, 1976), schemas (e.g. Fiske & Taylor, 1991) of group prototypes or stereotypes (e.g. Abrams & Hogg, 2010; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987; see Chapter 11). Behaviour that is consistent with our stereotypes or expectancies is attributed to stable internal factors, whereas expectancy-inconsistent behaviour is attributed to unstable or situational factors (e.g. Bell, Wicklund, Manko, & Larkin, 1976; Rosenfield & Stephan, 1977). When people explain behaviour that confirms their expectancy, they may simply rely on dispositions implied by a stereotype, with little or no effort to consider additional factors (Kulik, 1983; Pyszczynski & Greenberg, 1981) 2A selfesteem process: People’s need for secure self-esteem can be nurtured by making self-favouring comparisons between their ingroup and relevant outgroups. This process is a fundamental aspect of social identity theory (e.g., Tajfel & Turner, 1986; also Hogg & Abrams, 1988; see Chapter 11). Because people derive their social identity from the ‘groups to which they belong (a description and evaluation of themselves in terms of the defining features of the group), they have a vested interest in maintaining or obtaining an ingroup profile that is more positive than that of relevant outgroups. The ethnocentric attributional bias quite clearly satisfies this aim: it internally attributes good things about the ingroup and bad things about the outgroup, and it externally attributes bad things about the ingroup and good things about the outgroup. Attribution and stereotyping Societal and intergroup attribution processes significantly influence and are influenced by the stereotypes we have of groups in society: Stereotyping is not only an individual cognitive activity (see Chapter 2); it can also serve ego-defensive functions (making one feel good in contrast to others) and social functions (allowing one to fit in with other people’s world views) (Snyder & Miene, 1994) Groups invoke and accentuate existing stercotypes in order to attribute large-scale distressing events to the actions of specific outgroups ~ that is, scapegoats (Tajfel, 19812)SOCIAL KNOWLEDGE AND SOCIETAL ATTRIBUTIONS For instance, during the 1930s in Germany, the Jews were blamed for the economic crisis of the time. It was politically expedient to invoke the ‘miserly Jew’ stereotype to explain in sim- plistic terms the lack of money: there is no money because the Jews are hoarding it. Closer to home, stereotypes of immigrants as sponging on the state were invoked by “leavers” in the run-up to the June 2016 referendum that voted for Britain to leave the European Union. Stereotypes may also be invoked to justify actions committed or planned against an out- group (e.g, Crandall, Bahns, Warner, & Schaller, 2011). For instance, a group might develop a stereotype of an outgroup as dull-witted, simple, lazy and incompetent in order to explain. or justify the economic and social exploitation of that group. Social knowledge and societal attributions People do not wake up every morning and causally reconstruct their world anew: In general, wwe rely on well-learnt causal scripts (Abelson, 1981) and general eausal schemata. We stop, think and make causal attributions only when events are unexpected or inconsistent with expectations (e.g, Hastic, 1984; Langer, 1978; Pyszczynski & Greenberg, 1981), when we are ina bad mood (Bohner, Bless, Schwarz, & Strack, 1988), when we feel a lack of control (Liu & Steele, 1986) or when attributions are occasioned by conversational goals: for example, when we want to offer a particular explanation or justification of behaviour to someone (Hewstone & Antaki, 1988; Lalljce, 1981; Tetlock, 1983). Usually, we rely on a wealth of acquired and richly textured cultural knowledge that automatically explains what is going on around us, This knowledge resides in cultural beliefs, social stercotypes, collective ideolo- gies and social representations (see Box 3.4) Social representations ‘One way in which cultural knowledge about the causes of things may be constructed and transmitted is described by Moscovici’s theory of social representations (e.g. Farr & Moscovici, 1984; Lorenzi-Cioldi & Clémence, 2001; Moscovici, 1961, 1981, 1988; Purkhardt, Box 3.4 Your life ‘Avery strange custom: The cultural context of causal attribution 105 Social representations ‘rd.complex ahenomena thact familar and spe ‘form them into & Gun Semin tesa ficttious story about a Brazilian aborig- ine who visits Rio de Janeiro and then returns home to his tribe deep in the Amazonian rainforest to give an account ‘of the visit (Semin, 1980, p. 292). (On particular days more people than all those you have seen in your whole lifetime roam to this huge place of worship, an open hut the size of which you will never imagine. They come, chanting, singing, with symbols of their gods and once everybody is gathered the chanting drives away all alien spirits. Then, at the appointed time the priests arrive wearing colourful garments, and the chanting rises to war cries until three high priests, wearing black, arrive, All priests who were running around with sacred round objects leave them and at the order of the high priests begin the religious ceremony. Then, when the chief high priest gives a shrill sound from himself they all run after the single sacred round object that is left, only to kick it away when they get hold of it. Whenever the sacred object goes through one of the two doors and hits the sacred net the religious followers start to chant, pierc- ing the heavens, and most of the priests embark on a most ecstatic orgy until the chief priest blows the whis- tle on them. This is, of course, a description of a football match by someone who does not know the purpose or rules of the game, It illustrates an important point. For your explarta- tions to be meaningful they need to be grounded in a wider and more general interpretative framework that constitutes your socially acquired cultural knowledge.106 ‘CHAPTER 3 [ATTRIBUTION AND SOCIAL EXPLANATION, 1995). (See Chapter 5 for a discussion of the relationship between social representations and attitudes.) Social representations are understandings shared among group members. ‘They emerge through informal everyday communication. They transform the unfamiliar and complex into the familiar and straightforward, and they therefore provide a common-sense framework for interpreting our experiences. An individual or a specialist interest group develops a sophisticated, non-obvious, tech- nical explanation of a commonplace phenomenon (e.g. explaining mental illness in terms of biological or social factors rather than spiritual forces). This attracts public attention and becomes widely shared and popularised (i.e. simplified, distorted and ritualised) through informal discussion among non-specialists. It is now a social representation ~ an. accepted, unquestioned common-sense explanation that ousts alternatives to become the orthodox explanation. ‘Moscovici originally focused on the development of the theory of psychoanalysis, but his analysis is just as applicable to other formal theories and complex phenomena that have been transformed and simplified to become part of popular consciousness: for exam- ple, evolution, relativity, dietary and health theories, Marxism and climate change. The theory of social representations has come under some criticism, often for the rather smprecise way in which it is formulated (e.g. Augoustinos & Innes, 1990). Nonetheless, it does suggest how ordinary social interaction in society constructs common-sense or ‘naive’ causal theories that are widely used to explain events (Heider, 1958). As the world becomes increasingly complex the relevance of a social representations perspective becomes very appealing ~ for example, to help explain how the enormously complex dynamics surrounding the emergence and appeal of terrorist groups such as DAISH, Al-Qaeda and the Taliban is boiled down to a misleadingly simplistic conflict between Islam and the West. One source of criticism of social representations has been that itis difficult to know how to analyse social representations quantitatively. This problem has now largely been resolved. Appropriate quantitative techniques have been developed (Doise, Clemence, & Lorenzi- Cioldi, 1993), and practical descriptions of methodology have been published (Breakwell & Canter, 1993}, These methods include qualitative and quantitative analyses of interviews, questionnaires, observational dara and archival material. A good example of this methodo. logical pluralism is Jodelet’s (1991) classic description of social representations of mental illness in the small French community of Ainay-le-Chateau, in which questionnaires, inter: views and ethnographic observation were all used. Social representations, like norms (see Chapters 7 and 8), tend to be grounded in groups and differ from group to group such that intergroup behaviour can often revolve around a clash of social representations (Lorenzi-Cioldi &¢ Clémence, 2001). For example, in Western, countries, attitudes and behaviour that promote healthy lifestyles are associated with higher social status, and health promotion messages tend to come from middle-class professional groups (Salovey, Rothman, & Rodin, 1998). A social representations analysis suggests that, these messages are relatively ineffective in promoting healthy lifestyles for non-middle-class people because they are inconsistent with the wider representational framework of a good ife for such people. The European Union (EU) provides fertile ground for social representations research (e.g. Chryssochoou, 2000) that connects with the study of European identity dynamics (e.g Cinnitella, 1997; Huici, Ros, Cano, Hopkins, Emles, & Carmona, 1997). The EU is, in many ways, a prototypical social representation ~ a relatively new and technical idea that has its roots in complex economic matters such as free trade and subsidies. But the EU is now an accepted and commonplace part of European discourse which often emphasises more emo- tive issues of national and European identity ~ although the recent global and European economic and immigration crises have refocused attention on the nature of national borders and national identity and on economic and trade issues associated with the single currency and the concept of a European Central BankSOCIAL KNOWLEDGE AND SOCIETAL ATTRIBUTIONS Rumour and gossip Social representations are constructed in a way that resembles how rumours develop and are communicated (Allport & Postman, 1947; DiFonzo & Bordia, 2007). One of the earliest studies of rumour was conducted by Allport and Postman (1945), who found that if experi- mental participants described a photograph to someone who had not seen the photo, and then this person described it to another person, and so on, only 30 per cent of the original detail remained after five retellings. Allport and Postman identified three processes associ- ated with rumour transmission 1 Levelling — the rumour quickly becomes shorter, less detailed and less complex. 2 Sharpening — certain features of the rumour are sclectively emphasised and exaggerated, 3. Assimilation the rumour is distorted in line with people’s pre-existing prejudices, par- tialities, interests and agendas. ‘More naturalistic studies have found rather less distortion as a consequence of rumour transmission (e.g. Caplow, 1947; Schachter & Burdeck, 1955). ‘Whether or not rumours are distorted, and even whether rumours are transmitted at all, seems to depend on how anxious those who hear the rumour are (Buckner, 1965; Rosnow, 1980}. Uncertainty and ambiguity increase anxiety and stress, which lead people to seck out information to rationalise anxiety, which in turn enhances rumour transmission. (Check the third ‘What do you think? question. Here is one reason why Rajna wanted to )) Whether a rumour is distorted or becomes more precise depends on whether people approach the rumour with a critical or uncritical orientation. In the for mer case the rumour is refined, while in the latter (which often accompanies a crisis) the rumour is distorted. Rumours always have a source, and often this source purposely elaborates the rumour for a specific reason. The stock market is a perfect context for rumour elaboration ~ and, of course, the consequences for ordinary people’s everyday lives can be enormous. At the end of the 1990s, rumour played a significant role in inflating the value of ‘dot-com’ start-up companies, which then crashed in the NASDAQ meltdown early in 2000. More recently, there was enormous build-up and hype surrounding the launching of Facebook as a public company on the stock market in May 2012 ~ Facebook shares lost 25 per cent of their value in the two weeks following the launch, Rumour also played a significant role in the global stock market crash at the end of 2008 and beginning of 2009 (the market lost more than half its value), and in reports about Greek economic collapse that depressed the stock market in August 2011 and May 2012. ‘Another reason why rumours are purposely elaborated is to discredit individuals or groups. An organisation can spread a rumour about a competitor in order to undermine the competitor's market share (Shibutani, 1966}, or a group can spread a rumour to blame another group for a widespread crisis. A good example of this is the fabrication and prom- ulgation of conspiracy theories, which we discuss in the next subsection. But first, what about gossip? Gossip is informal talk, usually but not necessarily mali cious, behind the back of absent third parties (Foster, 2004; also see Baumeister, Zhang, & Vohs, 2004; Smith, 2014). In this respect itis narrower than rumour ~ rumour is about issues of significance to a group (a possible round of lay-offs) whereas gossip is about the personal characteristics of an absent other (a colleague's embarrassing sexual escapades) Gossip polices normative practices by vilifying those who violate norms; increases cohe- sion among those who are included in the circle of gossip; and empowers those who spread the gossip by making them appear to be ‘in the loop’, privy to secret information. and superior to the vietims of the gossip. In these respects, gossip serves a very clear social representational function, but of course, gossiping is also for many people just, great fun, pass a rumour on. 107108 CHAPTER3 ATTRIBUTION AND SOCIAL EXPLANATION Conspiracy theory n of widespread worn Conspiracy theories Conspiracy theories are simplistic and exhaustive causal theories that attribute widespread natural and social calamities to the intentional and organised activities of certain social groups that are seen as conspiratorial bodies set on ruining and then dominating the rest of humanity (Graumann & Moscovici, 1987). These groups are also perceived to be highly entitative (distinct, homogenous, inward looking ~ see Chapter 8), even cliquish (Grzesiak- Feldman & Suszek, 2008) One of the best-documented conspiracy theories is the myth, dating from the Middle Ages, of the Jewish world conspiracy (Cohn, 1966), which surfaces periodically and often results in massive systematic persecution. Other conspiracy theories include the belief that, immigrants are intentionally plotting to undermine the economy, that homosexuals are intentionally spreading HIV and that witches {in the Middle Ages) and DAISH (most recently) are behind virtually every world disaster you care to mention (e.g. Cohn, 1975) Research suggests that it is people who are personally willing to conspire who tend to endorse conspiracy theories most readily (Douglas & Sutton, 2011). Conspiracy theories wax and wane in popularity. They were particularly popular from the mid-seventeenth to the mid-cighteenth century: Everywhere people sensed designs within designs, cabals within cabals; there were court conspiracies, backstairs conspiracies, ministerial conspiracies, factional conspiracies, aristo- cratic conspiracies, and by the last half ofthe eighteenth century even conspiracies of gigan- tic secret societies that cut across national boundaries and spanned the Atlantic. ‘Wood (1982, p. 407) The accomplished conspiracy theorist can, with consummate skill and breath-taking ver- satility, explain even the most arcane and puzzling events in terms of the devious schemes and inscrutable machinations of hidden conspirators. Billig (1978) believes itis precisely this that makes conspiracy theories so attractive — they are incredibly effective at reducing uncer- tainty (Hogg, 2007b, 2012). They provide a causal explanation in terms of enduring disposi- tions that can explain a wide range of events, rather than complex situational factors that are less widely applicable. Furthermore, worrying events become controllable and easily remedied because they are caused by small groups of highly visible people rather than being due to complex sociohistorical circumstances (Bains, 1983} Not surprisingly, conspiracy theories are almost immune to disconfirming evidence. For example, in December 2006, the outcome of a three-year, 3.5-million-pound enquiry into the death in 1997 of Princess Diana was reported ~ although there was absolutely no evi- dence that the British Royal Family conspired with the British government to have her killed to prevent her from marrying an Egyptian Muslim, this conspiracy theory still persists There are also conspiracy theories about the 9/11 terrorist attacks in the United States in 2001 some Americans are absolutely convinced it was the doing of the US government, and in parts of the Muslim world, many people believe it was perpetrated by Israel (Lewis, 2004) A recent conspiracy theory has it that President Barack Obama is not only black, and pre- sumably not white, but not really an American at all! (See Box 3.5.) Societal attributions The emphasis on attributions as social knowledge surfaces in research on people’s explana: tions for large-scale social phenomena. In general, this research supports the view that causal attributions for specific phenomena are located within and shaped by wider, socially con- structed belief systems. For example, Catholics and non-Catholic Christians have subtly different attributional styles when explaining social phenomena and, in particular, the religious notion ofBox 3.5 Our world Barack Obama is black and not really an American Why is Barack Obama - the child of a Midwestern mother ‘white as milk’ and a Kenyan father ‘black as pitch’ (Obama, 2004, p. 10) - considered an African ‘American, but never Waite? Halberstadt, Sherman and Sherman (2011, p. 29) This isan example of hypodescent - a tendency to catego- rise children whose parents come from different status ‘groups, usually ethnic, into the subordinate group. Jamin Halberstadt and his colleagues have argued that hypodes- cent is a bias in the way we compare and classify features ‘of majority and minority group members, and the impor- tance that we give to distinctive features of the minority. (See Chapter 2 for a detailed discussion of salient or attention-capturing stimuli) In the case of Obama, the bias of hypodescent has been ‘elaborated into a conspiracy myth. In the United States, a SOCIAL KNOWLEDGE AND SOCIETALATTRIBUTIONS 109 full quarter of adult Americans, mainly right-wing social and religious conservatives, are ‘birthers: They believe that ‘Obama was not born in America and thus was ineligible to be president, and that there was a Democratic conspiracy ‘to conceal this. Even though there is overwhelming and incontroverti- bile proof that Obama was born in Hawaii (his official birth certificate was made public in 2008 and again in April 2011), birthers were not fazed. A'13 May 2011 Gallup poll showed that 23 per cent of Republican supporters remained birthers. Conspiracy theorists are tenacious. Asit became increasingly difficult to sustain the belief that ‘Obama was foreign-born, some birthers became ‘school- ers! who believe that because Obama is black there is no ‘way he could have gained entry to Harvard without cheat- ing and receiving special favours, and that - wait forit- the Democrats have a conspiracy going to conceal this as well ‘salvation’ (Li, Johnson, Cohen, Williams, Knowles, & Chen, 2014). Catholies are extrinsi- cally motivated and explain people’s behaviour, plight and ultimately salvation in terms of external factors (including ritual, the Pope, priests), whereas non-Catholic Christians are intrinsically motivated and proffer explanation in terms of internal factors in the form of, personally internalised religious tenets (e.g, the protestant work ethic ~ Furnham, 1984; Weber, 1930). Socioeconomic status and political ideology also influence attribution and social expla- nation. For example, research on explanations for poverty has shown that both the rich and the poor tend to explain poverty in terms of poor people's behaviour rather than the situa- tion that those people find themselves in (e.g. Feagin, 1972; Feather, 1974). This individualis- tic tendency is weaker among people with a more left-wing or liberal ideology and people living in developing countries where poverty is widespread (Pandey, Sinha, Prakash, & Tripathi, 198: Explanations for wealth tend to depend on political affiliation. In Britain, Conservatives often ascribe it to positive individual qualities of thrift and hard work, while Labour sup- porters attribute it to the unsavoury individual quality of ruthless determination (Furnham, 1983). Not surprisingly, there are also cross-cultural differences: for example, individualistic explanations are very common in Hong Kong (Forgas, Morris, & Furnham, 1982; Furnham & Bond, 1986) Similarly, explanations given for unemployment are influenced by people’s wider belief and value systems (Chapter 5). For example, Australian students preferred societal over indi- vidualistic explanations for unemployment; nominating defective government, social change and economic recession as more valid causes of unemployment than lack of motivation and personal handicap (Feather, 1985; see also Feather & Barber, 1983; Feather & Davenport, 1981). However, students who were politically more conservative placed less emphasis on societal explanations. Studies conducted in Britain show the same thing ~ societal explana- tions are more prominent than individualistic explanations, and that there is generalno ‘CHAPTER 3 [ATTRIBUTION AND SOCIAL EXPLANATION, agreement between employed and unemployed respondents (Furnham, 1982; Gaskell & Smith, 1985; Lewis, Snell, & Furnham, 1987}. Other research has focused on people's explanations for riots (social unrest, collective behaviour and riots are discussed in detail in Chapter 11). Riots arc enormously compli- cated phenomena where there are both proximal and distal causes ~ a specific event or action might trigger the riot, but only because of the complex conjunction of wider conditions. For instance, the proximal cause of the 1992 Los Angeles riots was the acquittal of white police officers charged with the beating of a black motorist, Rodney King (see Box 11.1}; however, this alone would have been unlikely to promote a riot without the background of racial unrest and socio-economic distress in the United States at the time. As with explanations of poverty, wealth and unemployment, people’s explanations for a specific riot are influenced by their sociopolitical perspective (e.g. Litton & Potter, 1985; Reicher, 1984, 2001; Reicher & Potter, 1985; Schmidt, 1972). Conservative members of the establishment tend to identify deviance, or personal or social pathology, while people with more liberal social attitudes tend to identify social circumstances. For example, Schmidt (1972) analysed printed media explanations of the spate of riots that occurred in American cities during 1967. The explanations could be classified on three dimensions: (a) legitimate—illegitimate, (b) internal-external cause, and (¢) institutional— environmental cause. The first two dimensions were strongly correlated, with legitimate external causes (e.g. urban renewal mistakes, slum conditions) going together and illegiti- mate internal causes (e.g. criminal intent, belief that violence works) going together. Media sources on the political right tended to identify illegitimate internal causes, whercas those classified as ‘left-centre’ (ie. liberal) emphasised legitimate external causes. Finally, Sniderman, Hagen, Tetlock and Brady (1986) investigated people’s explana tions for racial inequality and their preferences for different government policies. They used a national sample of whites in the United States (in 1972) and focused on the influ ence of level of education. They found that less-educated whites employed an ‘affect driven’ reasoning process. They started with their (mainly negative} feelings about blacks, then proceeded directly to advocate minimal government assistance. Having done this, they ‘doubled back’ to fill in the intervening link to justify their advocacy — namely that blacks were personally responsible for their own disadvantage. In contrast, better educated whites adopted a ‘cognition-driven’ reasoning process where they reasoned both forwards and backwards. Their policy recommendations were based on causal attributions for inequality, and in turn their causal attributions were influenced by their policy preference. Culture’s contribution Attribution and social explanation is not only affected by religious ideology, sociopolitical values, educational status, group membership and ethnicity, but also, perhaps unsurpris- ingly, by culture, People from different cultures often make very different attributions, make attributions in different ways or approach the entire task of social explanation in different ways (Chiu & Hong, 2007; Heine, 2016; Smith, Bond, & Kagitgibasi, 2006). Consequently, the potential for cross-cultural interpersonal misunderstanding is enormous. For example, the Zande people of West Africa have a dual theory of causality, where common-sense proximal causes operate within the context of witchcraft as the distal cause (Evans-Pritchard, 19375 see also Jahoda, 1979). This is, ironically, not really that different from moderate Christians’ belief in the proximal operation of scientific principles within the context of God as the distal cause. For the Zande, an internal-external distinction would make little sense.SOCIAL KNOWLEDGE AND SOCIETAL ATTRIBUTIONS 111 Another example: Lévy-Bruhl (1925) reported that the natives of Motumotu in New Guinea attributed a pleurisy epidemic to the presence of a specific missionary, his sheep, ‘wo goats and, finally, a portrait of Queen Victoria. Although initially quite bizarre, these sorts of attribution are easily explained as social representations. How much more bizarre are they than, for example, the postulation in physics of other universes and hypothetical particles shaped like strings or membranes as part of a unified theory to explain the origin, and structure of the cosmos (Hawking, 1988; Hawking & Mlodinow, 2010)? Horgan. exclaimed that “This isn’t physics any more. It’s science fiction with mathematics’ (Horgan, 2011, p. B7). One area of cross-cultural attribution research is the correspondence bias (discussed ear- lier in this chapter). We have seen that in Western cultures, people tend to make dispositional attributions for others’ behaviour (Gilbert & Malone, 1995; Ross, 1977), and that such dis- positional attributions become more evident over ontogeny (e.g. Pevers & Secord, 1973). In non-Western cultures, however, people are less inclined to make dispositional attributions (Carrithers, Collins, 8& Lukes, 1986; Morris & Peng, 1994). This is probably partly a reflec- tion of the more pervasive and all-enveloping influence of social roles in more collectivist non-Western cultures (Fletcher 8& Ward, 1988; Jahoda, 1982) and partly a reflection of a more holistic world view that promotes context-dependent, occasion-bound thinking (Shweder & Bourne, 1982). To investigate further the role of culture in dispositional attributions, Miller (1984) com- pared middle-class North Americans and Indian Hindus from cach of four age groups (adults, and 15-, 11- and 8-year-olds). Participants narrated prosocial and antisocial behav- iour and gave their own spontancous explanations of the causes of this behaviour. Miller coded responses to identify the proportion of dispositional and contextual attributions that participants made. Among the youngest children there was little cross-cultural difference (sce Figure 3.8). As age increased, however, the two groups diverged, mainly because the Americans increasingly adopted dispositional attributions. For context attributions, the results were reversed. Culture and attribution Isthe puppet responsible forts own ctions? Easerners are Jess likely than Westerners to make disposttional atvibutons about people ~ let alone Puppets!ATTRIBUTION AND SOCIAL EXPLANATION Figure 3.8 Dispositional attributions as a function of age and cultural background ans and indian Hi of dispositional a he age of 15, there isa clear differen strengthens in adultnood, vi an indians in their North Americans Indian Hindus 02 on Proportion of disposi tially do not differ in the butions made for behaviour that ans being significantly enburions Byeas Tiyears I5years Adult Age Ame! ‘The important lesson this study teaches us is that cultural factors have a significant impact on attribution and social explanation. (We return to the role of culture in social behaviour in Chapter 16.) a, Summary ) J ‘© People are naive psychologists seeking to understand the causes of their own and other people's behaviour ‘© Much like scientists, people consider consensus, consistency and distinctiveness information in deciding whether to attribute behaviour internally to personality traits and dispositions, or exter- nally to situational factors. ‘© The attributions that we make can have a profound impact on our emotions, self-concept and relationships with others, There may be individual differences in propensities to make internal or ‘external attributions. ‘© People are actually poor scientists when it comes to making attributions. They are biased in many different ways, the most significant of which are a tendency to attribute others’ behaviour disposi- tionally and their own behaviour externally, and a tendency to protect the self-concept by exter- nally attributing their own failures and internally attributing their successes. ‘© Attributions for the behaviour of people as ingroup or autgroup members are ethnocentric and based on stereotypes. However, this bias is affected by the real or perceived nature of intergroup relations, ‘© Stereotypes may originate in a need for groups to attribute the cause of large-scale distressing ‘events to outgroups that have (stereotypical) properties that are causally linked to the events, ‘© People resort to causal attributions only when there is no readily available social knowledge (e.g scripts, causal schemata, social representations, cultural beliefs) to explain things automatically ‘© Social representations are simplified causal theories of complex phenomena that are socially con- structed through communication contextualised by intergroup relations. Rumour and gossip may play a key role in social representations,LITERATURE, FILM ANDTV 113 "© Conspiracy theories are one particularly bizarre but sadly prevalent type of causal theory that ‘often persis in the face of overwhelming evidence that the t theory is wrong. ‘+ People’s world views and identity in society (. religion, wealth, politics, culture) significantly impact how they make attributions and explain socal phenomena (e.g. poverty, unemployment, rots) Actor-observer effect Discount Attribution Distinctiveness infor Attributional style Essentialsm Belief in ajust world Ethnocentrism ‘Causal schemata Cognitive miser Consensus information Consistency information Conspiracy theory Correspondence bias Correspondent inference Covariation model False consensus effec Hedonic relevance Illusion of control Internal (or dsposit Level of explanation External (or situational) attribution Fundamental atribution error Intergroup attributions Motivated tactician Naive psychologist (or scientist) Non-common effects ‘Outcome bias Personalism Self handicapping Selt-perception theory Self-serving bias Social representations Social identity theory Stereotype Uimate attribution error mation ct jonal) attribution JFK ‘A.1991 film by Oliver Stone. It stars Kevin Costner as a New Orleans district attorney who reopens the case to find out who really assassinated JFK on 22 November 1963, in Dallas, and what the process/plot behind it was. This is a wonderful encounter with conspiracy the: and people's need to construct a causal explanation, how- ever bizarre, of a disturbing event. The film also stars Tommy Lee Jones and Sissy Spacek. Going Clear: Scientology and the Prison of Belief ‘A multi-award-winning 2015 documentary directed by ‘Alex Gibney, based on an earlier 2013 book by Lawrence Wright, which, with the aid of archive footage, dramatic reconstructions and interviews describes the origins, his- tory and nature of L. Ron Hubbard's Church of Scientology. The film documents the awful lengths to which a group can go to protect its ideology and world view - any divergence is seen as heresy or blasphemy, and is severely and crvelly punished in order to make sure that everyone believes in the group's explanation of the nature of things. ‘Macbeth Shakespeare's 1606-7 tragedy in which three witches prophesise a string of evil deeds committed by Macbeth during his bloody rise to power, including the murder of the Scottish king Duncan. The causal question is whether the prophecy caused the events - or whether was there some other complex of causes. For those of you who pre- fer films, Justin Kurzel has directed a highly acclaimed 2015 film version of Macbeth that stars Michael Fassbender in the title role and Marion Cotillard as Lady Macbeth Legally Blonde ‘A 2001 award-winning comedy directed by Robert Luketic and starring Reese Witherspoon. Witherspoon plays Elle Woods, a stereotypically breathless self-confident blonde southern California sorority girl. This sounds pretty much ‘one of a million such films, but this one is actualy funny,114 CHAPTER3 ATTRIBUTION AND SOCIAL EXPLANATION relatively clever and has more going on. Itis a nice vehicle for exploring the way that people construct someone's personality from the way they appear and behave, and then it can be difficult for the target to break free of the pigeonhole. Elle, like most people, is more complex and less superficial than her appearance and some of her behaviour lead one to think. But as she tres to be taken seriously as a law student and a person, she finds that those around her continually construct her personality on the basis of superficial cues. Guided questions J 11 What is meant by locus of control? How does locus of control affect the way we invoke effort, abil- ity fate and chance to explain behaviour, and how might ths influence our own success in life? 2. Do attribution processes create problems in close relationships, or vice versa? 3 Sometimes our mental short-cuts lead us into error, One such error is the correspondence bias. What is this bias, how i it produced and how can it be combatted? 4. What is meant by self-handicapping? Provide a real-world setting in which it can be applied 5 The term conspiracy theory has entered everyday language. Can social psychology help us under- stand what purpose these theories serve, and even combat them? Qn, Learn more ) Fiske, ST, & Taylor, 5. E. (2013). Social cognition: From brains to culture (2nd ed)) Los Angeles: Sage. Most recent edition of Fiske and Taylor's classic social cognition text itis comprehensive, detailed and well written, and it also covers the recent developments in social neuroscience, Fletcher, G., & Fincham, FD. (Eds) (1991). Cognition in close relationships. Hillsdale, NJ: Erlbaum. A collection of leading scholars contributes detailed chapters on attribution and other sociocog- nitive approaches to close relationships. Hewstone, M.(1989). Causal attribution: From cognitive processes to collective beliefs. Oxford: Blackwell ‘A-comprehensive and detailed coverage of attribution theory and research, which also includes coverage of European perspectives that locate attribution processes in the context of society and intergroup relations. Hilton, D. J. (2007). Causal explanation: From social perception to knowledge-based causal atribu- tion. In A. W. Kruglanski & ET. Higgins (Eds), Social psychology: Handbook of basic principles (2nd ced, pp. 232-253). New York: Guilford. A comprehensive coverage of research on causal attribu- tion processes and social explanation, Macrae, C.N, & Quadflieg, S. (2010). Perceiving people. In ST Fiske, D.. Gilbert, &G. Lindzey (Eds), Handbook of social psychology (Sth ed. Vol. 1, pp. 428-463). New York: Wiley. Comprehensive coverage of what we know about person perception ~ how we form and use our cognitive repre- sentations of people. Moskowitz, G. B. (2005), Social cognition: Understanding self and others. New York: Guilford. A rela tively recent comprehensive social cognition text that is written in an accessible style as an intro- duction to the topicSmith, E.R, 1994). Social cognition contributions to attribution theory and research. In P.G. Devine, DL Hamilton, & TM. Ostrom (Eds), Social cognition: Impact on social psychology (pp. 77-108). San Diego, CA: Academic Press, A focused coverage of social cognitive dimensions of attribution processes Trope. Y, & Gaunt, R. (2007). Attribution and person perception. In M. A. Hogg, & J. Cooper (Eds), The SAGE handbook of social psychology: Concise student edition (pp. 176-194). London: SAGE. A relatively recent, comprehensive and above all readable overview of attribution research Weary, G, Stanley, M.A., & Harvey). H. (1989). Attribution. New York: Springer-Verlag. A discussion of applications of attribution theory and the operation of attribution processes in clinical settings and everyday life outside the laboratory. LEARN MORE us