Jussim's main thesis is that much research in social psychology has overemphasised error while ignoring accuracy in social perception (Jussim Reference Jussim2012). Jussim's eloquent argument is squarely aimed at social psychologists, and he clearly articulates methodological and interpretation issues with often cited studies in favour of error and self-fulfilling prophecies. He also highlights general research issues that are relevant across science, such as careful research design, the importance of replications (or lack thereof), and the need to focus on effect sizes rather than statistical significance. Indeed, it is the small effect sizes found across studies and meta-analyses (Chapters 6–9) that should make it difficult for any reasonable scientist to persist with the notion that error dominates accuracy and that self-fulfilling prophecies have, on average, powerful effects on other's behaviour. Likewise, the accuracy found across studies (Chs. 17 and 18) makes it clear that accuracy is real and worth studying.
What Jussim is arguing against largely stems from deep rooted ideology in social psychology (as he discusses in Chs. 2 and 10). In other areas of science, however, the argument that social perception should be accurate would not be considered controversial and, in fact, may be taken as an assumption. Specifically, evolutionary or biological approaches examine behaviour in many different species addressing how these behaviours are adaptive and functional. Of course, while it is naive to assume that all behaviour is adaptive, it would be surprising to expect error to dominate accuracy from this view point.
For any organism, the fundamental problems are survival and reproduction, and these often encompass navigating a social world in which individuals, for example, compete, cooperate, and find a mate in a pool of other individuals. Consequently, non-human animals demonstrate a variety of adaptations to assess the behaviour of others and there is a large literature concerning the evolution of animal signals used to communicate, among other things, behaviour (Krebs & Dawkins Reference Krebs, Dawkins, Krebs and Davies1984). For example, in antagonistic encounters with other individuals of the same species, the primary decision to be made is to fight or not. Given the potential costs, injury or even death, we might expect that animals will possess perceptual/cognitive adaptations to assess the risks by assessing fighting ability in their opponents (Enquist & Leimar Reference Enquist and Leimar1983; Parker Reference Parker1974). Indeed, there is evidence that animals such as mice and crabs make adaptive decisions about fighting based on the assessment of the relative fighting abilities of their opponents (L. M. Gosling et al. Reference Gosling, Atkinson, Dunn and Collins1996; Hazlett Reference Hazlett1996). Accuracy could arise because specific traits of some species can be related to fighting success. For example, variable black facial patterns in paper wasps are related to body size and social dominance (Tibbetts & Dale Reference Tibbetts and Dale2004), and in gelada baboons high status males have the reddest chests (Bergman et al. Reference Bergman, Ho and Beehner2009). Individuals could base their decisions to fight on appearance linked cues to fighting ability allowing them to compete when likely to win and to avoid costly agonistic interactions when likely to lose.
An evolutionary view then has a prediction concerning accuracy and inaccuracy in social perception because this view tends to assume that perception serves an adaptive function: The external world is full of information that can be used to guide adaptive and functional behaviours (Zebrowitz-McArthur & Baron Reference Zebrowitz-McArthur and Baron1983). If, in our evolutionary past, information were presented about a person's behaviour (e.g., likelihood of cooperation or aggression) in any way, then an advantage would accrue to those who utilised these cues and those individuals would leave more genes behind in the next generation. An individual may not last long if they make too many errors in important social domains and that individual may not leave many offspring compared to an individual who is able to more accurately predict the behaviour of others. Of course, this does not mean stimulus-perception links should be innate, selection pressures could favour accuracy or adaptive behaviour via learning or calibration mechanisms.
Other researchers have emphasised that social perception is functional rather than error prone. The evolutionary view has much in common with an ecological approach (Gibson Reference Gibson1979), which has an emphasis that perception is for doing. In fact, Gibson's approach highlights the utility of perception, including in the social domain, and he noted that some human behaviour is of critical importance to survival and reproduction and people should pay special attention to them – factors such as sex, fighting, and cooperation. Indeed, one aspect of an evolutionary approach is that we might predict highest accuracy in judgements that are most relevant in evolutionary terms. For example, alongside accuracy in terms of personality attribution, an evolutionary approach also highlights accuracy for judgements of traits such as cooperation, aggression, current health, and sexual behaviour.
In terms of accuracy, clearly, some stereotypes are accurate. Like the questions at the start of Ch. 15, I ask my class each year to imagine a bank robber and overwhelmingly they imagine men, not women, and are accurate because bank robberies are mainly committed by men. Jussim presents a convincing case for the accuracy of stereotypes about groups and also notes the accuracy of judgements of individuals. In support of his argument for accuracy there is an increasingly large literature demonstrating such accurate judgements from minimal information. Early research using a “zero acquaintance” paradigm (Ch. 20) in which unacquainted individuals without interaction rated each other on personality traits demonstrated self-other correspondence for various personality traits (Albright et al. Reference Albright, Kenny and Malloy1988; Kenny et al. Reference Kenny, Albright, Malloy and Kashy1994; Passini & Norman Reference Passini and Norman1966). Using minimal information, self-other correspondence has also been found when observers rate static face photographs for personality (Little & Perrett Reference Little and Perrett2007) and when guessing political inclinations (Rule & Ambady Reference Rule and Ambady2010). Accuracy is not limited to judging the person themselves either: judges can accurately infer some personality traits from brief viewing of targets’ bedrooms and offices (S. D. Gosling et al. Reference Gosling, Ko, Mannarelli and Morris2002).
There is mounting evidence that people are somewhat accurate in judging the behaviour of others from minimal information. These studies include accuracy in assessing traits that are especially important in terms of evolutionary pressures. Studies have shown accurate social perception relevant to: (1) Mate choice, such as accuracy in judging inclination for short-term mating (Boothroyd et al. Reference Boothroyd, Jones, Burt, DeBruine and Perrett2008) and health and stress judgements (Little et al. Reference Little, McPherson, Dennington and Jones2011); (2) Choosing who to ally with, such as accuracy in judging cooperation (Little et al. Reference Little, Jones, DeBruine and Dunbar2013) and deception (Bond et al. Reference Bond, Berry and Omar1994); and (3) Fighting ability, such as accuracy in judging who would win in physical fights (Little et al. Reference Little, Trebicky, Havlicek, Roberts and Kleisner2015; Trebicky et al. Reference Trebicky, Havlicek, Roberts, Little and Kleisner2013). Of course, individual accuracy for these traits is not perfect, and in fact is often only just greater than chance. Given the minimal information, and the often noisy measures involved (e.g., using questionnaire measures), this accuracy still seems impressive.
Evolutionary inspired work on accuracy has also examined how facial features may relate to real behaviour through links to hormones that regulate facial growth. For example, facial masculinity positively relates to testosterone level (Penton-Voak & Chen Reference Penton-Voak and Chen2004) and also relates to perceived dominance (Perrett et al. Reference Perrett, Lee, Penton-Voak, Rowland, Yoshikawa, Burt, Henzi, Castles and Akamatsu1998) and physical strength (Fink et al. Reference Fink, Neave and Seydel2007). This hormone-appearance link may then provide a link between appearance and actual dominant behaviour (Mazur & Booth Reference Mazur and Booth1998). Indeed, facial masculinity also predicts risk-taking in financial decision tasks (Apicella et al. Reference Apicella, Dreber, Campbell, Gray, Hoffman and Little2008) and chess games (Dreber et al. Reference Dreber, Gerdes, Graensmark and Little2013), which is consistent with some aspects of the effects of testosterone on behaviour (Mazur & Booth Reference Mazur and Booth1998).
Another example comes from facial width to height ratio (fWHR), a ratio which is also associated with testosterone in men (Lefevre et al. Reference Lefevre, Lewis, Perrett and Penke2013). fWHR predicts self-reported and other rated dominance (Mileva et al. Reference Mileva, Cowan, Cobey, Knowles and Little2014) and is associated with perceived aggression (Carre et al. Reference Carre, Morrissey, Mondloch and McCormick2010), trait dominance using questionnaires, and real aggression in a naturalistic setting (Carre & McCormick Reference Carre and McCormick2008; Carre et al. Reference Carre, McCormick and Mondloch2009). These hormone–appearance–behaviour links provide an interesting and biologically informed source of accuracy in social perception.
Alongside accuracy, an evolutionary approach also provides a framework for understanding low accuracy in social perception in the form of theories concerning animal signals (Krebs & Dawkins Reference Krebs, Dawkins, Krebs and Davies1984). In this framework, receivers are under selection to predict the behaviour of signallers while signallers are also under selection to manipulate receivers and so their signals may be honest or dishonest (see e.g., Krebs & Dawkins Reference Krebs, Dawkins, Krebs and Davies1984). An honest signal contains accurate information about the signaller and a dishonest or deceptive signal contains inaccurate information. Under these pressures, we can expect honest signals to evolve where accuracy leads to a benefit for both signallers and receivers, and dishonest signals to evolve where inaccuracy benefits signallers but not receivers. Such reasoning may help explain why low accuracy is seen for some social perception. For example, in situations of detecting lying, the liar does not want to be caught and so we would expect lower accuracy. On the other hand, people often want to display, for example, their social skills to others, and so, such judgements may be more accurate. While people may want to exaggerate aspects such as social skill, honesty in the signal may come about because deviations from reality are hard to fake, making such signals honest.
Evolutionary theory may even help explain some biases in social perception. Error management theory (Haselton & Buss Reference Haselton and Buss2000) is an evolutionary perspective in which bias can arise when there are differential costs and benefits to overestimation or underestimation of traits in others. One example has been termed the sexual over-perception bias (Haselton Reference Haselton2003), in which men and women face different costs and benefits to casual sexual encounters. The chance of pregnancy places more potential cost for women in casual sexual encounters than men, while men gain the opportunity to increase the number of genes they pass on at relatively low cost. Given this difference in cost/benefit, we can expect men and women to differ in their perception of sexual intent of the opposite sex. Specifically, selection would favour men who did not miss out on sexual opportunities, and so we expect men to be more likely to interpret women's behaviour as interested in sex, even when they are not. The cost of the “miss” is low while the benefit of the “hit” is high for men, whereas these costs are different for women. This bias does appear to be apparent with men's estimates of women's sexual intent being higher than women's (see Haselton [2003] for brief review). This is then an error, but an error that could serve an adaptive function. Adaptive function in social perception can then also encompass the achievement of goals, not just accuracy (Zebrowitz-McArthur & Baron Reference Zebrowitz-McArthur and Baron1983).
To conclude, the idea of judging an individual's behaviour accurately may be seen as inherently undesirable, but this in no way implies that it is not important to attempt to understand this area and Jussim makes a pervasive argument that accuracy is important. In fact, the evidence that people appear to make such judgements based on minimal information despite society's discouragement implies that this is an area of fundamental importance in social perception. An evolutionary approach can provide important insights into social perception, such as:
-
1. Accuracy should be expected because selection pressures will have shaped social perception to be functional.
-
2. We should expect greatest accuracy in social domains that are important for survival and reproduction.
-
3. We should expect lowest accuracy when it is adaptive for the observed to hide their behaviour from observers.
-
4. Bias and inaccuracy are not always irrational if they serve the interest of the perceiver.
Overall, it is clear that there is much accuracy in social perception, even when observers have minimal information. Given the complexity of behaviour, it is unlikely that observers will be perfectly accurate but an evolutionary view strongly predicts that people will behave as rational observers and so in many cases social perception should favour adaptive responses.
Jussim's main thesis is that much research in social psychology has overemphasised error while ignoring accuracy in social perception (Jussim Reference Jussim2012). Jussim's eloquent argument is squarely aimed at social psychologists, and he clearly articulates methodological and interpretation issues with often cited studies in favour of error and self-fulfilling prophecies. He also highlights general research issues that are relevant across science, such as careful research design, the importance of replications (or lack thereof), and the need to focus on effect sizes rather than statistical significance. Indeed, it is the small effect sizes found across studies and meta-analyses (Chapters 6–9) that should make it difficult for any reasonable scientist to persist with the notion that error dominates accuracy and that self-fulfilling prophecies have, on average, powerful effects on other's behaviour. Likewise, the accuracy found across studies (Chs. 17 and 18) makes it clear that accuracy is real and worth studying.
What Jussim is arguing against largely stems from deep rooted ideology in social psychology (as he discusses in Chs. 2 and 10). In other areas of science, however, the argument that social perception should be accurate would not be considered controversial and, in fact, may be taken as an assumption. Specifically, evolutionary or biological approaches examine behaviour in many different species addressing how these behaviours are adaptive and functional. Of course, while it is naive to assume that all behaviour is adaptive, it would be surprising to expect error to dominate accuracy from this view point.
For any organism, the fundamental problems are survival and reproduction, and these often encompass navigating a social world in which individuals, for example, compete, cooperate, and find a mate in a pool of other individuals. Consequently, non-human animals demonstrate a variety of adaptations to assess the behaviour of others and there is a large literature concerning the evolution of animal signals used to communicate, among other things, behaviour (Krebs & Dawkins Reference Krebs, Dawkins, Krebs and Davies1984). For example, in antagonistic encounters with other individuals of the same species, the primary decision to be made is to fight or not. Given the potential costs, injury or even death, we might expect that animals will possess perceptual/cognitive adaptations to assess the risks by assessing fighting ability in their opponents (Enquist & Leimar Reference Enquist and Leimar1983; Parker Reference Parker1974). Indeed, there is evidence that animals such as mice and crabs make adaptive decisions about fighting based on the assessment of the relative fighting abilities of their opponents (L. M. Gosling et al. Reference Gosling, Atkinson, Dunn and Collins1996; Hazlett Reference Hazlett1996). Accuracy could arise because specific traits of some species can be related to fighting success. For example, variable black facial patterns in paper wasps are related to body size and social dominance (Tibbetts & Dale Reference Tibbetts and Dale2004), and in gelada baboons high status males have the reddest chests (Bergman et al. Reference Bergman, Ho and Beehner2009). Individuals could base their decisions to fight on appearance linked cues to fighting ability allowing them to compete when likely to win and to avoid costly agonistic interactions when likely to lose.
An evolutionary view then has a prediction concerning accuracy and inaccuracy in social perception because this view tends to assume that perception serves an adaptive function: The external world is full of information that can be used to guide adaptive and functional behaviours (Zebrowitz-McArthur & Baron Reference Zebrowitz-McArthur and Baron1983). If, in our evolutionary past, information were presented about a person's behaviour (e.g., likelihood of cooperation or aggression) in any way, then an advantage would accrue to those who utilised these cues and those individuals would leave more genes behind in the next generation. An individual may not last long if they make too many errors in important social domains and that individual may not leave many offspring compared to an individual who is able to more accurately predict the behaviour of others. Of course, this does not mean stimulus-perception links should be innate, selection pressures could favour accuracy or adaptive behaviour via learning or calibration mechanisms.
Other researchers have emphasised that social perception is functional rather than error prone. The evolutionary view has much in common with an ecological approach (Gibson Reference Gibson1979), which has an emphasis that perception is for doing. In fact, Gibson's approach highlights the utility of perception, including in the social domain, and he noted that some human behaviour is of critical importance to survival and reproduction and people should pay special attention to them – factors such as sex, fighting, and cooperation. Indeed, one aspect of an evolutionary approach is that we might predict highest accuracy in judgements that are most relevant in evolutionary terms. For example, alongside accuracy in terms of personality attribution, an evolutionary approach also highlights accuracy for judgements of traits such as cooperation, aggression, current health, and sexual behaviour.
In terms of accuracy, clearly, some stereotypes are accurate. Like the questions at the start of Ch. 15, I ask my class each year to imagine a bank robber and overwhelmingly they imagine men, not women, and are accurate because bank robberies are mainly committed by men. Jussim presents a convincing case for the accuracy of stereotypes about groups and also notes the accuracy of judgements of individuals. In support of his argument for accuracy there is an increasingly large literature demonstrating such accurate judgements from minimal information. Early research using a “zero acquaintance” paradigm (Ch. 20) in which unacquainted individuals without interaction rated each other on personality traits demonstrated self-other correspondence for various personality traits (Albright et al. Reference Albright, Kenny and Malloy1988; Kenny et al. Reference Kenny, Albright, Malloy and Kashy1994; Passini & Norman Reference Passini and Norman1966). Using minimal information, self-other correspondence has also been found when observers rate static face photographs for personality (Little & Perrett Reference Little and Perrett2007) and when guessing political inclinations (Rule & Ambady Reference Rule and Ambady2010). Accuracy is not limited to judging the person themselves either: judges can accurately infer some personality traits from brief viewing of targets’ bedrooms and offices (S. D. Gosling et al. Reference Gosling, Ko, Mannarelli and Morris2002).
There is mounting evidence that people are somewhat accurate in judging the behaviour of others from minimal information. These studies include accuracy in assessing traits that are especially important in terms of evolutionary pressures. Studies have shown accurate social perception relevant to: (1) Mate choice, such as accuracy in judging inclination for short-term mating (Boothroyd et al. Reference Boothroyd, Jones, Burt, DeBruine and Perrett2008) and health and stress judgements (Little et al. Reference Little, McPherson, Dennington and Jones2011); (2) Choosing who to ally with, such as accuracy in judging cooperation (Little et al. Reference Little, Jones, DeBruine and Dunbar2013) and deception (Bond et al. Reference Bond, Berry and Omar1994); and (3) Fighting ability, such as accuracy in judging who would win in physical fights (Little et al. Reference Little, Trebicky, Havlicek, Roberts and Kleisner2015; Trebicky et al. Reference Trebicky, Havlicek, Roberts, Little and Kleisner2013). Of course, individual accuracy for these traits is not perfect, and in fact is often only just greater than chance. Given the minimal information, and the often noisy measures involved (e.g., using questionnaire measures), this accuracy still seems impressive.
Evolutionary inspired work on accuracy has also examined how facial features may relate to real behaviour through links to hormones that regulate facial growth. For example, facial masculinity positively relates to testosterone level (Penton-Voak & Chen Reference Penton-Voak and Chen2004) and also relates to perceived dominance (Perrett et al. Reference Perrett, Lee, Penton-Voak, Rowland, Yoshikawa, Burt, Henzi, Castles and Akamatsu1998) and physical strength (Fink et al. Reference Fink, Neave and Seydel2007). This hormone-appearance link may then provide a link between appearance and actual dominant behaviour (Mazur & Booth Reference Mazur and Booth1998). Indeed, facial masculinity also predicts risk-taking in financial decision tasks (Apicella et al. Reference Apicella, Dreber, Campbell, Gray, Hoffman and Little2008) and chess games (Dreber et al. Reference Dreber, Gerdes, Graensmark and Little2013), which is consistent with some aspects of the effects of testosterone on behaviour (Mazur & Booth Reference Mazur and Booth1998).
Another example comes from facial width to height ratio (fWHR), a ratio which is also associated with testosterone in men (Lefevre et al. Reference Lefevre, Lewis, Perrett and Penke2013). fWHR predicts self-reported and other rated dominance (Mileva et al. Reference Mileva, Cowan, Cobey, Knowles and Little2014) and is associated with perceived aggression (Carre et al. Reference Carre, Morrissey, Mondloch and McCormick2010), trait dominance using questionnaires, and real aggression in a naturalistic setting (Carre & McCormick Reference Carre and McCormick2008; Carre et al. Reference Carre, McCormick and Mondloch2009). These hormone–appearance–behaviour links provide an interesting and biologically informed source of accuracy in social perception.
Alongside accuracy, an evolutionary approach also provides a framework for understanding low accuracy in social perception in the form of theories concerning animal signals (Krebs & Dawkins Reference Krebs, Dawkins, Krebs and Davies1984). In this framework, receivers are under selection to predict the behaviour of signallers while signallers are also under selection to manipulate receivers and so their signals may be honest or dishonest (see e.g., Krebs & Dawkins Reference Krebs, Dawkins, Krebs and Davies1984). An honest signal contains accurate information about the signaller and a dishonest or deceptive signal contains inaccurate information. Under these pressures, we can expect honest signals to evolve where accuracy leads to a benefit for both signallers and receivers, and dishonest signals to evolve where inaccuracy benefits signallers but not receivers. Such reasoning may help explain why low accuracy is seen for some social perception. For example, in situations of detecting lying, the liar does not want to be caught and so we would expect lower accuracy. On the other hand, people often want to display, for example, their social skills to others, and so, such judgements may be more accurate. While people may want to exaggerate aspects such as social skill, honesty in the signal may come about because deviations from reality are hard to fake, making such signals honest.
Evolutionary theory may even help explain some biases in social perception. Error management theory (Haselton & Buss Reference Haselton and Buss2000) is an evolutionary perspective in which bias can arise when there are differential costs and benefits to overestimation or underestimation of traits in others. One example has been termed the sexual over-perception bias (Haselton Reference Haselton2003), in which men and women face different costs and benefits to casual sexual encounters. The chance of pregnancy places more potential cost for women in casual sexual encounters than men, while men gain the opportunity to increase the number of genes they pass on at relatively low cost. Given this difference in cost/benefit, we can expect men and women to differ in their perception of sexual intent of the opposite sex. Specifically, selection would favour men who did not miss out on sexual opportunities, and so we expect men to be more likely to interpret women's behaviour as interested in sex, even when they are not. The cost of the “miss” is low while the benefit of the “hit” is high for men, whereas these costs are different for women. This bias does appear to be apparent with men's estimates of women's sexual intent being higher than women's (see Haselton [2003] for brief review). This is then an error, but an error that could serve an adaptive function. Adaptive function in social perception can then also encompass the achievement of goals, not just accuracy (Zebrowitz-McArthur & Baron Reference Zebrowitz-McArthur and Baron1983).
To conclude, the idea of judging an individual's behaviour accurately may be seen as inherently undesirable, but this in no way implies that it is not important to attempt to understand this area and Jussim makes a pervasive argument that accuracy is important. In fact, the evidence that people appear to make such judgements based on minimal information despite society's discouragement implies that this is an area of fundamental importance in social perception. An evolutionary approach can provide important insights into social perception, such as:
1. Accuracy should be expected because selection pressures will have shaped social perception to be functional.
2. We should expect greatest accuracy in social domains that are important for survival and reproduction.
3. We should expect lowest accuracy when it is adaptive for the observed to hide their behaviour from observers.
4. Bias and inaccuracy are not always irrational if they serve the interest of the perceiver.
Overall, it is clear that there is much accuracy in social perception, even when observers have minimal information. Given the complexity of behaviour, it is unlikely that observers will be perfectly accurate but an evolutionary view strongly predicts that people will behave as rational observers and so in many cases social perception should favour adaptive responses.