Towards a Sociology of Attitudes

by David Voas
University of Essex

Sociological Research Online, 19 (1) 12
<http://www.socresonline.org.uk/19/1/12.html>
10.5153/sro.3289

Received: 17 Jun 2013     Accepted: 10 Dec 2013    Published: 28 Feb 2014

Abstract

The psychology of attitudes has made important contributions to knowledge over many decades, but it has tended to cause confusion in the study of social attitudes. From a sociological perspective, attitudes should be defined as prescriptive or evaluative judgements, not as individual predispositions to act in particular ways or to view something favourably or unfavourably. The basic distinction that has been lost is that between personal preferences, tastes or feelings on the one hand and social attitudes on the other. Attitudes also need to be distinguished from beliefs and values. Contrary to the current consensus in psychology, if not actual practice in sociology, attitudes are more usefully regarded as observable characteristics than as latent constructs.

This perspective is applied to survey questions related to attitudes and then both theoretically and empirically to the concepts of tolerance and trust, with the aim of illustrating what might emerge from a more sociological treatment of attitudes. The analysis involves cross-national comparisons using the 2008 European Values Study. Distaste for multiple kinds of out-groups seems to be a better predictor of not wanting a homosexual as a neighbour than disapproval of homosexuality, and similar results are found for other 'others'. With trust, the evidence shows that not only individual but also national differences in the perceived fairness of others have substantial effects. To the extent that trust is rooted in a moral evaluation of people in general, it qualifies as an attitude itself rather than being just a belief about the attitudes of others.

Keywords: Beliefs, Preferences, Values, Behaviour, Tolerance, Trust

Introduction

1.1 Social norms shape and are shaped by what individuals regard as good or bad. Many surveys focus on these evaluative judgements, loosely referred to as social attitudes. One might therefore expect to find a well developed body of sociological theory to explain what attitudes are and how they function in society. In fact it was psychologists who colonized the field nearly a century ago; sociologists have neglected these issues, although they have given some attention to the gap between what we say and what we do (Deutscher 1973).

1.2 Psychological definitions and theories of attitudes have proliferated, particularly since the Second World War. In keeping with the person-centred focus of the discipline, they have emphasized the subjectivity of attitudes. To say that an attitude is 'a psychological tendency that is expressed by evaluating a particular entity with some degree of favour or disfavour' (Eagly & Chaiken 1993: 1) puts the emphasis on the internal mental state. These accounts make common cause with emotivism, the idea in moral philosophy (popularised by AJ Ayer and CL Stevenson in the 1930s and 40s) that statements about right and wrong are merely expressions of personal approval or disapproval. The psychology of attitudes typically fails to distinguish, or at least blurs the distinction, between individual preferences and prescriptive judgements, that is between what I like and what I think that people ought to do. In addition, the theories do not always clearly demarcate between matters of fact and taste.

1.3 The psychology of attitudes has produced valuable spin-offs, especially in the measurement of attitudes and in investigating their reciprocal connections to beliefs, emotion and behaviour. What it has not done is to provide a good basis for the study of attitudes in society, a situation recognised even by some psychologists in this field. An 'outside observer would be surprised to find little emphasis on the structure of the social environment relative to the considerable emphasis on psychological structure and processes' (Eagly & Chaiken 1993: 682).

1.4 The problem is deeper than simply needing to think more about how norms interact with individual inclinations, however. I shall argue that we ought to see attitudes as normative statements about the social order rather than as subjective expressions of individual likes or dislikes. For sociologists and political scientists, attitudes are social phenomena that emerge from but are not reducible to the inner workings of human minds. The argument is not that psychologists should redefine attitudes (though there is a good deal of conceptual confusion surrounding the topic), but that social scientists who discuss them in these terms are guilty of a category mistake (Ryle 1949). From a sociological perspective, attitudes should be defined as prescriptive evaluations, not as predispositions to respond subjectively in a favourable or unfavourable way.

Defining social attitudes

2.1 My objective is to formalize the concept of 'attitude' as it is (or ought to be) used by sociologists. In doing so it is important to show how attitudes relate to feelings, beliefs, values and behaviour. An attitude is an everyday judgement, a normative view on a specific matter. It offers an evaluation rather than a descriptive statement of fact, and implicitly or explicitly involves notions such as good/bad, right/wrong, ought/ ought not. An attitude may be expressed or unexpressed; it may motivate or be influenced by behaviour, but it is basically propositional.

2.2 Attitudes, like some values, are normative; they represent evaluations. Values, though, are abstract; one could think of them as latent constructs that are manifested in the specific judgements we call attitudes. Value theorists seek to identify the key dimensions along which individual priorities can be located, with examples being traditional versus secular-rational values and survival versus self-expression (Inglehart & Welzel 2005) or conservation versus openness to change and self-enhancement versus self-transcendence (Schwartz 2007). More specific values might include concepts such as fairness and loyalty (Graham et al. 2013) or benevolence and achievement (Schwartz 2007). In contrast to these general principles, attitudes refer to concrete issues, for example whether state welfare provision is too high or too low.

2.3 The basic distinction that has been lost in classical treatments of attitudes is that between personal preferences, tastes or feelings on the one hand and social attitudes on the other. Social attitudes are prescriptive statements (about what people ought to do) or evaluative judgements that are intended to be general (about what is good or bad, right or wrong). It is important to recognise that people make two different kinds of judgements. Some are personal choices or purely subjective assessments; others express standards that apply collectively. It is not necessary for anyone to believe that there are objective, universal or absolute standards in ethics or aesthetics (though ordinary people are much less likely than sociologists to be cultural relativists); what matters is that we recognise the difference between what we like to do and what we think that everyone should do.

2.4 Some examples may help. The statements 'I don't like being asked by that charity to give up my time' or 'I'd rather watch a DVD than go to the cinema' express personal preferences; they do not prescribe how anyone else should feel. The statements 'Everyone should do voluntary work for good causes' or 'It's sad that people sit around at home watching DVDs rather than going out' express judgements that apply to all of us. Of course it may not be obvious what is being asserted; if Sally says 'Going to the cinema is a waste of time', we may want to press her on whether she is describing her own feelings or making a general claim about what people should do.

2.5 Even at the level of individual psychology it is doubtful that social attitudes should be treated as a species of personal preferences, but in any event they certainly need to be separated by sociologists. Societies rest on a foundation of shared norms. I am perfectly happy for my neighbour to eat liquorice rather than chocolate, though to my mind such tastes are perverse. I regard it as wrong for him to play loud music at 2 a.m., even if I personally have no objection. One of the interesting aspects of social change is the tendency of some matters to migrate from being matters of public morality to matters of personal choice while others go in the opposite direction. Premarital and same-sex relationships are now widely accepted as being no business of anyone else's, while 60 years ago they were subject to social sanctions; 20 years ago most Europeans would have been astonished by the idea that smoking in a bar was anything other than an individual decision.

2.6 Just as social attitudes are not matters of individual taste, neither do they concern matters of fact: descriptive statements that can be true or false. Although it is subject to continuing debate, the distinction between facts and values has been central to philosophical thinking since the Enlightenment. For example, 'Mother Teresa was only concerned about others' is descriptive, and so either true or false; 'Mother Teresa was an example to us all' is prescriptive, and so right or wrong. 'Smoking causes cancer' is almost certainly true, but it does not imply the attitude that smoking should be banned.

2.7 Note that evaluative judgements (like the statement about Mother Teresa) do not have to include words like 'good' or 'should', and conversely those terms can be used in a descriptive context. For instance, the terms 'good driver' or 'good musician' are factual descriptions meaning simply that the person demonstrates a certain skill. By contrast, 'good person' is an evaluative judgement, unless one has previously specified that a good person is simply someone who does X, Y and Z.

2.8 This last example points to a complication: the existence of 'thick concepts' (Williams 1985). Some terms may be viewed as descriptive but carry an evaluative or normative force, e.g. considerate, promise, lie, courage, coward, brutality, treachery. The descriptive and evaluative dimensions can be considered separately, however. The descriptive statements are true or false, and the judgement depends on the existence of a community with particular values. We can refer explicitly to 'technical terms' if we wish to avoid the normative connotations (so that in the sociology of religion, for example, terms such as 'apostasy', 'cult' and 'religiosity' are used descriptively and do not carry the pejorative force that is common in ordinary language).

2.9 The psychology of attitudes has tended to treat descriptive and normative statements in the same way, where both are at least potentially evaluative. Thus the statement 'A Rolls-Royce is a fine example of automotive craftsmanship' expresses a positive view about that make of car. It is no wonder, then, that 'attitudes' measured in such ways often seem to be so weakly correlated with behaviour. It is perfectly possible to believe this statement and many others that highlight desirable features of the Rolls-Royce without having the slightest interest in owning or driving one. Moreover, the fact that I can think that the car is a marvellous piece of engineering while having no interest in driving one is not even particularly interesting: there are many unsurprising reasons for such a disjunction between evaluation and behaviour (e.g. beliefs about its cost, a preference for smaller vehicles, attitudes about ostentation and environmental impact). By contrast, while it is entirely possible for attitudes as defined here to be inconsistent with behaviour - I might state that everyone should do more to save fuel while driving a Rolls - in such cases the contrast could be worth investigating.

2.10 As mentioned in the example above, 'people should do more voluntary work' is an attitude. Someone holding this opinion may or may not be a volunteer, and it is not helpful to define an attitude as 'a predisposition to behave in a particular way' (Procter 2008: 208). The relationship between evaluation or prescription and action is an empirical matter, and here sociologists have made important contributions. In a classic study (in which the direction of social desirability bias comes as a surprise today), American hotel and restaurant staff almost never refused to serve a Chinese couple despite endorsing a exclusionary policy (LaPiere 1934; see Deutscher 1973 for an overview of subsequent research).

Departures from the psychological tradition

3.1 Attitudes have been defined in many ways by social psychologists since the first half of the 20th century. Although the debate continues, the emerging consensus in that discipline is that attitudes are latent constructs. They are both formed by, and manifested through, beliefs, feelings, and behaviour. They are dispositions (either enduring or transient) to evaluate something in a particular way. For a discussion and defence of this definition, see Eagly & Chaiken (1993: 1-17; 2005: 744-9).

3.2 The conceptualisation of social attitudes outlined above is different. In the first place, I suggest that it is more useful from a sociological perspective to regard attitudes as evaluative judgements, rather than as mental states that give rise to those judgements. (This view can also be found in psychology; see for example Kruglanski & Stroebe 2005: 324). They are entities on the same level as beliefs, feelings and behaviour, and so are directly measurable (albeit with error and the usual problems of validity and reliability). Attitudes may indeed be both consequences and causes of beliefs, feelings and behaviour, but they can be observed and identified independently of them. Finally - and this is the crux of the matter - the standard psychological treatments have completely missed the defining feature of social attitudes. Attitudes are evaluative judgements applying to others as well as ourselves. They are not merely personal likes and dislikes; they concern how people in general ought to think, feel and behave.

3.3 The identification of the judgements themselves as attitudes avoids the complexities of deciding whether latent constructs are states, tendencies or dispositions, how they exist prior to being crystallised, the role of memory, and other issues about the nature of inner mental states. Such technical problems lead to abstruse descriptions that are challenging to grasp and apply, e.g. 'attitudes are emergent properties of the activity of microconceptual networks that are potentiated by contextually situated objects, goals and task demands' (Bassili & Brown 2005: 552).

3.4 The characterisation offered here is narrower as well as simpler than that found in psychology. It limits the scope of the term to normative and prescriptive assessments. Subjective preferences along the lines of 'I like chocolate' are not usefully regarded as social attitudes; they are simply expressions of personal taste. Likewise evaluations that are not value judgements - that is to say, that are factual rather than normative (e.g. 'swimming is a good form of exercise') - are not worth classifying as attitudes. Attitudes are opinions, but not all opinions are attitudes: not, in particular, those that concern matters of taste or empirical questions.

3.5 This narrowing of scope is consistent with a long-term trend in social psychology. Various authors have noted that the early researchers in the field (including Gordon Allport) defined attitudes so broadly that the concept encompassed virtually any mental disposition to behave in particular ways, thereby laying claim to the entire discipline. Over time the need to measure attitudes and to distinguish between different types of motivation led to greater specificity (Krosnick et al. 2005: 22). Restricting the term to genuinely social judgements is a natural step.

The origin or justification of attitudes

4.1 To recap, attitudes are statements of value: claims about good/bad, right/wrong, should/shouldn't. Although such claims cannot be derived directly from descriptive statements, matters of fact can provide reasons for making normative or prescriptive statements. There is an implicit syllogism in which the factual statement serves as one premise. For example:

4.2 Attitudes may be driven by beliefs, feelings, values or indeed behaviour in some combination. Some survey questions are not about attitudes, but about the beliefs (and sometimes the feelings or preferences) that underpin them. The intention is to understand which beliefs might lead to, or at least be used to justify, particular attitudes. If we ask respondents whether 'immigrants drive down wages', it is not to test their knowledge of labour market economics but because such beliefs might be linked to attitudes about immigration.

4.3 Social scientists have long been aware that just as beliefs and feelings can underpin attitudes, the possession of certain attitudes may influence beliefs and feelings. 'Wishful thinking' and 'positive thinking' - believing what one wants to be true - are examples, but there are many others. (The psychological research on the tendency to make beliefs congruent with attitudes is particularly valuable; a good overview can be found in Marsh & Wallace 2005.) Conversely, although the usual expectation is that attitudes cause behaviour, the influence can run in the other direction when we modify our attitudes to justify our behaviour.

4.4 The extent to which attitudes do in fact lead to behaviour is a standard topic for research. Mounting evidence by the late 1960s that attitudes were often only weakly related to behaviour caused something of a crisis for psychological theorists, but these 'conclusions did not come as a surprise to sociologists who had questioned the importance of personal dispositions and had emphasized instead social context and norms as determinants of human action' (Ajzen & Fishbein 2005: 175). The degree to which attitudes are associated with action is of interest to sociologists, but it is an empirical question rather than a postulate that defines the field.

4.5 Whereas attitudes are elicited directly in conversation, in interviews, or from self-completed questionnaires, values are latent constructs and to some degree abstract. They are general principles about what matters, and attitudes will be among their overt manifestations. Values are conceptions of what is desirable and important: they are ends, not means. Values are general rather than specific, so that support for equality could be a value, for example, whereas support for particular policies intended to promote equality would be attitudes. Values help to shape both preferences and attitudes and so indirectly motivate behaviour. They provide criteria for judgement and implicitly set priorities.

4.6 How values are structured and organised is a matter for theory; see for example Inglehart and Welzel (2005), Schwartz (2012), Graham et al. (2013). In the top left part of Table 1, the ten values in the Schwartz schema have been arranged in a 2x2 table where the key distinctions are between their personal or social focus on the one hand and their connection with protection or progression on the other. One point to note is that values need not be pro-social; some priorities will be individual, like power or hedonism. It is common to adopt a social focus in discussing values, but many of our ends are personal.

4.7 The final column lists the moral foundations identified by Haidt; they correspond reasonably closely with the socially focused Schwartz values. Inglehart's value system does not match as closely, though one could argue that his secular-rational versus traditional dimension is connected to the personal versus social focus distinction, while the survival versus self-expression dimension correlates with the protection versus progression scale.

4.8 The last row shows preferences and attitudes positioned in relation to personally and socially orientated values respectively. Values in the social realm are likely to be expressed as attitudes; values in the personal sphere will emerge as preferences.

Examples - the British Social Attitudes survey 2010

5.1 I have argued that the characterisation of attitudes described above is more useful to sociologists than the main alternatives from psychology. A further defence of this approach is that the concept is already used in the way described. One simple form of demonstration comes from a review of the questionnaires used for the British Social Attitudes (BSA) survey, which has run annually since 1983 (with gaps only in 1988 and 1992). It is a high quality sample survey conducted by NatCen Social Research using a random sample of more than 3,000 adults every year. In addition to a relatively stable core of questions, various different topics are covered in depth each year (often as the result of funding from government departments or academic research projects).

5.2 The BSA 2010 dataset available from the UK Data Archive includes 878 variables. Some are derived from other variables, and many are descriptive of the respondents (e.g. age, sex, education, etc.), but most contain the answers to specific questions on the topics covered by the survey. In 2010 the main areas included social security, health, transport, housing, and school choice. There is a lengthy face-to-face interview with every respondent, in addition to which a self-completion questionnaire is left to be returned by post.

5.3 As argued above, social researchers may wish to ask about beliefs and preferences as well as attitudes. In investigating how people think that the environment should be managed, for example, it is relevant to find out how much they know about climate change. Nevertheless, if the characterisation of attitudes offered here is appropriate, the large majority of the questions should concern normative or prescriptive judgements. An examination of the questionnaire confirms that they do. A typical question on welfare, for example, is:

Imagine an unmarried couple who split up. They have a child at primary school who remains with the mother. Do you think that the father should always be made to make maintenance payments to support the child?

5.4 Many questions are not so clear cut, however, and it is worth considering several examples in detail.

If there were a general election tomorrow, which political party do you think you would be most likely to support?

Some questions can be hard to classify. The question 'Which political party do you prefer?' seems to be framed in relation to preferences, but many people regard their political choices as attitudes: I see my support for party X not as a matter of taste but as a consequence of the opinion that party X has the best policies, that people ought to support party X, and so on. Similarly the question above seems to be framed in relation to belief (about one's own future behaviour), but it basically seeks to elicit party support via a hypothetical situation. Strictly speaking the behaviour (voting choice) may not be determined by the attitude (party choice), but in practice it is sufficiently hard to think of factors that would override the political attitude that it can be inferred from the intended behaviour.

5.5 Would you say that someone in Britain was or was not in poverty if . . .
. . . they had enough to buy the things they really needed, but not enough to buy the things most people take for granted?

This question might be taken in a literal sense to be about definitions: what counts as poverty? In practice, though, 'poverty' is a 'thick concept' of the sort mentioned earlier; the descriptive classification has prescriptive implications. Choosing one threshold rather than another leads to different decisions about the appropriate level of benefit payments, for example. While one could try to separate the definitional issues from the normative consequences, this question appears on the survey because what counts as poverty is not merely a semantic issue. Respondents are being asked about right or wrong ways of deciding who is poor.

5.6 How much do you agree or disagree with this statement. . .
There are not enough playgrounds for young children in my local area?

This issue might be either a matter of fact or a matter of value, depending on the degree of consensus about how much is 'enough'. Given general agreement that one needs X square metres in playground space per Y children, or playgrounds no further than Z metres from any home, the question is then a factual one about whether the criteria are satisfied in a particular area. In reality, however, the number of playgrounds needed is probably a matter of opinion on which there are many different views. That being so, the question solicits an attitude about whether there ought to be more (according to the 'right' standards, as opposed to 'the' standards).

5.7 All in all, how satisfied or dissatisfied would you say you are with the way in which the National Health Service runs nowadays?

This question is another attitudinal item masquerading as something else. While it appears to ask about the respondent's feelings (of satisfaction or otherwise), the management of the NHS is unlikely to be regarded as a matter of taste. The underlying question is whether or not the NHS is well run. Our evaluative judgement is general, being one that we would expect other right-minded people to share. (We leave aside the possibility that someone could be satisfied with a badly run service if they never use it and do not pay for it.) As with the playground item, there might be a fact/value issue: if there is a consensus about standards then whether those standards are met may be an empirical question. Lacking such a consensus, and indeed in the absence of expert knowledge about what standards are reasonable and what results are achieved, the opinion expressed is probably attitudinal. It might be pure affect (in psychological terms): 'I like the NHS'. More probably it represents an evaluation like 'The NHS is good/bad' or 'The people responsible for the NHS do / do not do the right things'.

5.8 Now suppose you wanted to see a GP about a bad chest infection that was stopping you from doing the things you normally do. How long do you think it would be reasonable to have to wait for an appointment to see a GP about this?

Although this question is framed in personal terms, it is clearly asking about what would be reasonable for anyone. There is an implicit trade-off between cost and performance. Economists sometimes seek to quantify preferences by asking people how much they would pay to do X or avoid Y; similarly, this question asks the respondent to identify the level at which the cost of waiting becomes unacceptable. Although preferences may differ, the notion of what is reasonable makes it plain that the answer applies to everyone.

5.9 Overall, how satisfied or dissatisfied are you with your local area as a place to live?

This question does seem to relate to personal feelings rather than an evaluation that would hold for everyone. Naturally one's satisfaction or dissatisfaction may stem from features that one regards as normatively good or bad, but here the issue seems more subjective. Many reasons for liking or disliking an area are a matter of taste: urban versus rural, the presence or absence of personal connections, and so on.

5.10 How much do you agree or disagree with each of these statements. . .
I worry a lot about the standard of living I will have when I reach retirement age?

At first glance the question seems to be about feelings (being worried), but it might more accurately be seen as one about belief: will you have enough to live on in old age? Of course one might believe the answer to be 'no' and still not worry, or vice versa, which suggests that the question is flawed if it was intended to elicit expectations.

5.11 How much do you trust British governments of any party to place the needs of the nation above the interests of their own political party?

And how much do you trust politicians of any party in Britain to tell the truth when they are in a tight corner?

The concept of trust will be discussed in some detail below. I will argue that trust is a belief about someone else's attitudes and their likelihood of being guided by them. Where the subject endorses the attitude in question, the belief implies an evaluative judgement of the person being trusted. If we think that governments should place the needs of the nation above those of the party, believing that they would do otherwise amounts to believing that they do wrong, which is itself an attitude.

Attitudes and tolerance

6.1 Attitudes do not exist in isolation. As discussed above, they can be influenced by beliefs and preferences, and that influence is often reciprocal. In the next two sections I shall examine the complex relationships underlying the key concepts of tolerance and trust. These case studies help to show the utility of distinguishing between feelings, attitudes, values and beliefs.

6.2 As a starting point, the discussion above helps us to clarify concepts such as prejudice, stereotyping and intolerance. Prejudice is a feeling, a propensity to like or dislike someone triggered purely because they possess a particular characteristic. A stereotype is the belief that someone with a particular characteristic has various other qualities. Tolerance is an attitude, a willingness to adapt to something that one dislikes. (Tolerating something one likes or to which one is indifferent is cost-free and requires no prescriptive push.)

6.3 Tolerance is an example of a two-place relation: should X be allowed to do Y? (By contrast, feelings involve a unitary object: one likes or dislikes X.) To tolerate means being willing to permit X to do Y, but the focus is typically on people, not on actions. I will therefore avoid using the term 'tolerate' to apply to behaviour (for example drug use or abortion). It is reasonable to ask 'are we willing to allow Y (for any X)?', but in this context we are interested in distinctions between different types of X.

6.4 Tolerance comes into play when feelings and values press in opposite directions. It takes no effort to tolerate someone we like who is doing things we approve of; it seems inappropriate even to use the notion of toleration in such circumstances. Tolerance becomes an issue when we are faced with the question of whether or not to give rights to a type of person we dislike. Should an inebriated visitor be tolerated or shown to the door? Should an extremist be allowed to address a particular meeting? Should a released prisoner be welcomed into the neighbourhood? An interesting question is how prejudice (a positive or negative feeling based on knowing a particular characteristic of the individual) is related to (in)tolerance (an attitude concerning what certain kinds of people should or shouldn't be allowed to do).

6.5 Previous scholarly work on tolerance is largely found in the political science literature, where the focus is on the willingness to grant civil rights to people who are disliked. Note that the 'allowing X to do Y' structure leads to measurement problems that have been recognised for some time. Standard batteries confound beliefs and feelings about X with the issue about rights. If Americans are increasingly willing to allow books by communists into school libraries, for example, it may be because they are now less hostile to communists than in the past rather than because they have become more supportive of free expression. To put the matter in the terms used here, what has changed are the beliefs about and the feelings towards communists, not necessarily the scope of values.

6.6 In sociology, the focus is commonly on prejudice: feelings towards a whole group of people. One approach to measurement attempts to isolate feelings about X (groups) from variations in attitudes about Y (situation). For example, the 'feeling thermometer' is sometimes used in surveys (including the British Social Attitudes survey in 2008, and the US General Social Survey with some regularity), where the choice of a number between 0 and 100 indicates cool (negative) or warm (positive) feelings about a group of people such as Hindus, convicted criminals, etc.

6.7 A classical way of measuring tolerance relates to social distance. It is explicitly a two-place relation: 'would you be willing to accept X as a Y?' In particular, the Bogardus social distance scale uses points defined as follows: being willing to accept a member of the group named (Hindus, convicted criminals, etc.) as

6.8 The roots of tolerance or intolerance can be well hidden. Consider a specific attitude: 'we should continue to allow Somali refugees to settle in Britain.' One might disagree with this statement because of feelings about Somalis (e.g. racial prejudice, distaste for their culture or religion) or about immigration in general (e.g. that changes to the status quo are unwelcome). To that extent intolerance is a natural consequence of negative feelings. The anti-immigration attitude might alternatively be based on beliefs, however. For example, one might believe that Somalis have few useful skills or little interest in integrating, or that immigrants take jobs from natives or depress wage rates. In such cases a lack of tolerance is not necessarily connected to feelings. Values might also play a part, to the extent that they lead to the view that (all else being equal) people should be allowed free movement, or the disadvantaged should be helped, or native interests should be safeguarded, or people with different values are a threat, and so on. For various reasons, then, attitudes (e.g. intolerance of Somali immigration) may not depend on feelings (e.g. prejudice against Somali immigrants).

6.9 The challenge with attitudes is to understand how they originate, and tolerance is a good case in point. One important question is whether people are intolerant in specific cases because they disapprove of particular characteristics or because they generally dislike out-groups. For example, people might be intolerant of homosexuals, immigrants or drug addicts because they disapprove of homosexuality, immigration or drug use respectively, or alternatively they might simply dislike 'others' in general. Intolerance could be specific or generalized.

6.10 Analysis of the European Values Study (EVS) may provide some answers. The EVS is a large-scale, cross-national, longitudinal survey research programme on the beliefs, preferences and attitudes of residents of every country in Europe, very broadly defined. The project began in 1981, when a thousand citizens in each of the European member states of that time were interviewed. Every nine years, the survey has been repeated in an increasing number of countries. The fourth wave in 2008/2009 covered no fewer than 46 countries from Iceland to Azerbaijan, with a minimum sample size of 1,500 in each. In total there are nearly 70,000 people in the dataset. As Great Britain and Northern Ireland, and East and West Germany, were covered separately, there are 48 territories in all.

6.11 Respondents were given a list of 15 groups of people (including 'drug addicts', 'homosexuals', 'immigrants/foreign workers', etc.) and asked to identify any that they would not, generally speaking, like to have as neighbours. If tolerance were mostly about the specific group and activity, then one might expect attitudes to that group and activity to be more important than any general disposition to like or dislike people who are different. A simple test can be carried out using binary logistic regression, where the dependent variable is the willingness to have a homosexual neighbour. There are two independent variables: the answer to a question about whether homosexuality can be justified (on a scale from 1 to 10), and the total number of other groups of people not liked as neighbours. The full set of groups in the order listed on the questionnaire is: people with a criminal record, people of a different race, left wing extremists, heavy drinkers, right wing extremists, people with large families, emotionally unstable people, Muslims, immigrants/foreign workers, people who have AIDS, drug addicts, homosexuals, Jews, Gypsies, and Christians. The count of groups not liked excludes homosexuals (the category in question) and Christians (because in most countries they do not constitute an out-group). More fully specified models were also tried, but virtually all of the effects of socio-demographic controls are mediated by these two variables.

6.12 As shown in Table 2, reluctance to have members of various other out-groups as neighbours is a better predictor of not wanting to live near homosexuals than attitudes to homosexuality. There are similar findings for immigrants and drug use. Thus it appears that generalised intolerance (not wanting members of out-groups as neighbours) is a better predictor of specific intolerance (not wanting someone with a particular characteristic as a neighbour) than disapproval of the characteristic itself. Or to put it another way, intolerance in this instance is likely to be the result of a general distaste for people who are different rather than an assessment of the situation in the light of relevant beliefs and values.

Table 2. Binary logistic regression of 'would not like a homosexual as a neighbour' on general intolerance and specific disapproval of homosexuality


6.13 The exceptions (at the bottom of the list) are places in which the number of not-liked neighbours has been depressed by a 'tick all that apply' (TATA) format. Research has shown that when presented with this sort of battery, respondents select more items when asked separately about each in turn than if they are requested to tick all that apply in a list. There are various explanations: the yes/no format takes - and allows - more time for reflection; the TATA format leads to satisficing behaviour; and (in this case) when it is socially undesirable to display intolerance, it is easier with TATA to avoid mentioning all answers that actually apply. Unfortunately the battery was administered in different ways in different countries.

6.14 Intolerance seems to exist in a generalised form applying to multiple out-groups. It seems plausible that generalised intolerance also applies to multiple contexts (e.g. degrees of social distance or varieties of civil rights), but this hypothesis needs to be tested.

6.15 To summarize, a specific attitude - for example that women in the armed forces should or should not be allowed to serve in combat - may be based on values related to liberty, equality, authority, tradition, purity, and so on, but it may also be based on (or merely defended by) various beliefs and feelings about female soldiers, combat, the importance of one or the other, or the consequences of adopting a particular policy. The elements of tolerance include beliefs and feelings about X and Y, the perceived significance of Y, and the weight given to different values that may support one conclusion or the other. The extent to which intolerance is a product of prejudice or something else is an empirical question.

6.16 Research that recognises the distinctions between feelings, beliefs, values and attitudes can help us to understand the sources of both specific and generalised intolerance. If promoting tolerance is a policy objective, sociological research may point to which strategy is most likely to succeed. Values may be important, but feelings are obviously relevant (e.g. hostility towards others), as are beliefs (for example about the benefit or harm of difference and change). Opinion leaders and policy makers who want to change behaviour have a greater chance of doing so if they know how certain attitudes are produced. Is it a matter of changing beliefs (and if so which ones), or altering feelings, or inducing a re-ordering of values?

Attitudes and trust

7.1 In the previous section we considered the case of a feeling (prejudice) that was relevant to an attitude (tolerance). With trust the connection is different. Trust, I argue, is a belief about the attitudes of others. We rely on others to behave in ways that we expect on the basis that they hold particular attitudes and will act in accordance with them. I trust strangers on the street not to harm me because I believe that they would consider such behaviour wrong. I trust my mother with my wallet because I believe that she would regard it as wrong to take anything from it. I may not trust my mother to avoid embarrassing me in front of my friends if I know that she does not think it wrong to open photo albums, tell stories about my past, and so on. If the context is general rather than specific - as with generalized trust where we refer to unknown actors doing unspecified things - then trust amounts to the evaluation that people are generally good, fair, or at least benign. Trust, then, is an attitude itself as well as a belief about the attitudes of others.

7.2 How much I trust X to do Y depends on two things. One is the strength of my belief about X's relevant attitudes; the other is my assessment of the strength of X's attitude. I may not know Joe well enough to judge what he thinks that he should or shouldn't do, or I may not be sure how strongly he adheres to the attitudes that I rely on. One way we make decisions is to look for signals (Bacharach & Gambetta 2001). Rightly or wrongly I may be more inclined to trust a stranger who is wearing a suit than one who is wearing a hoodie: I have different degrees of confidence that their attitudes and actions will be benign.

7.3 I have described trust as implying the belief that others will act in accordance with the attitudes we believe that they hold. This extra assumption is needed to deal with intentional wrongdoing. A thief may believe that theft is wrong. Trust therefore involves a belief not only that the other person shares some key values but also that the attitudes on which we rely are sufficiently strong not to be easily overridden by self-interest.

7.4 Trust involves a belief in someone else's honesty, fairness or good intentions; on the basis of this belief, one exposes oneself to a risk of loss. Uslaner (2002) argues that trust is connected to values as much as beliefs. He distinguishes between strategic trust, which is based on experience, and moralistic trust, which is not. 'Moralistic trust is a moral commandment to treat people as if they were trustworthy' (Uslaner 2008: 101). Interestingly, he then continues 'The central idea behind moralistic trust is the belief that most people share your fundamental moral values.' These two principles seem rather different. Believing that people share your values is similar to believing that they are good people who will not harm you. That is not the same as being trusting because you want to be a good person and to trust is the right thing to do.

7.5 Of course Uslaner agrees that knowledge can matter a great deal in particular instances: we may be reluctant to entrust a valuable possession to someone we know has an expensive drug habit. Where we have no knowledge of specific individuals, belief in the predictability of the goodwill of strangers is still relevant. People in cities lock their doors; there is ample evidence that failure to be careful might be punished, the costs of being careful are low, and the benefits of being trusting are also low. Trust in these circumstances is still a strategic matter, with risks and rewards being weighed up.

7.6 It is possible to investigate empirically the connection between trust and the belief that other people will behave well. The generalized trust question that has been used for decades in survey research is:

Generally speaking, would you say that most people can be trusted or that you can't be too careful in dealing with people?

That question is included in the 2008 EVS, accompanied by another that directly addresses the issue of beliefs about the attitudes of others:

Do you think that most people would try to take advantage of you if they got the chance, or would they try to be fair? Where would you place your view on this scale? [from 1 to 10]

7.7 Somewhat surprisingly the responses to the two questions are not especially highly correlated at the individual level. The within-country correlation coefficients range from 0.17 in Moldova to 0.54 in Belarus, with the average for the 47 countries included being 0.34. (Azerbaijan is omitted because the data for these two questions appear to be unreliable; trust is unexpectedly high while perceived fairness is exceptionally low.) In aggregated form, however, the association between the two variables is clear. Figure 1 shows the relationship between the perceived fairness of others and willingness to trust at a national level.

7.8 Countries vary substantially in how far one can rely on other people and in whether they can be trusted. One question that arises is whether the relationship between the beliefs that most people try to be fair and that most people can be trusted is the same cross-nationally. In other words, are Cypriots less trusting than Danes purely because they expect to encounter exploitation rather than fairness, or are there other factors at work? It is easy to establish that the cross-national differences in trust are not explained purely by corresponding differences along the 'others would take advantage versus others are fair' scale. On average, untrusting people in Denmark have the same level of belief in the fairness of others as trusting people in Cyprus. (Both are exactly 7 on the 1-10 scale.)

Figure 1. Perceived fairness of others and generalized trust by country, 2008

7.9 One way of illustrating the differences is shown in Figure 2. The countries in the EVS dataset have been grouped by similarity in the response profiles for these two questions. (For the sake of clarity not all national groups are included in the graph.) If the relationship between perceived fairness and trust was the same across Europe, all of the curves would be identical, and countries would vary only in the relative frequencies of people who had low or high trust. In fact, though, the patterns are different.

Figure 2. Generalized trust by perceived fairness of others, for country groups, 2008

7.10 The country groups are intended merely to illustrate different patterns; the names are used for convenience and may not apply to all members of the group:

Across most of the European continent the belief that other people are inclined to take advantage of you, given the opportunity, is associated with a low propensity to trust. Scandinavia is exceptional in that even people who have little confidence in the fairness of others still show comparatively high levels of trust. (There are fewer people at this end of the scale in Scandinavia than elsewhere.) Trust rises steeply beyond the midpoint of the perceived fairness scale, as one would expect. Nevertheless, in many countries (represented here by the loosely labelled 'Catholic' and 'Ottoman' categories) only a minority of even those respondents who think that other people would try to be fair give a positive answer to the generalized trust question.

7.11 One possible explanation is that 'most people can be trusted' and 'you can't be too careful' are actually separate dimensions, so that when respondents are forced to choose they might favour caution to trust (Miller & Mitamura 2003, Yamagishi et al. 1999). Especially if the belief that others would try to take advantage is nationally widespread, one might decide that being careful is more important than being trusting.

7.12 The hypothesis is that for individuals who believe that others try to be fair, their likelihood of being trusting is strongly influenced by the national level of perceived fairness. Holding perceived fairness constant, people who live in countries where it is commonly thought that others will take advantage are more likely to choose 'you can't be too careful' in answering the generalized trust question. This hypothesis is supported by the data. As shown in Table 3, there is a clear association between trust and the national mean on the exploitation versus fairness scale, even among individuals who strongly endorse the idea that people in general try to be fair. The correlation between the first two columns of figures is 0.59. (The correlation between perceived fairness and trust overall, the first and third columns, is 0.85, as illustrated in Figure 1.)

Table 3. National variations in trust among people who believe that others try to be fair


7.13 There are some exceptions; in Greece and Macedonia, for example, both perceived fairness and trust are low, but the (relatively few) people who believe in the fairness of others tend to be trusting. Conversely Russia and France score well on perceived fairness but badly on trust. Cultural differences in the importance attached to caution could be responsible. And despite the enormous care taken with translation of survey questionnaires, it is hard not to worry about the problems of finding exact equivalents for the colloquial nuances in expressions such as 'you can't be too careful' or 'most people would try to take advantage of you if they got the chance'.

7.14 The findings from these analyses are consistent with the view that the generalized trust question measures caution as well as trust, which would help to explain why the association with perceived fairness is relatively weak at the individual level and varies substantially at the national level. They might also be seen as supporting the argument that what produces trust is not adult experience of civic participation, but rather childhood socialization into the value of having faith in others (Uslaner 2002).

7.15 The cross-national differences also underline the point that reliance based on coercion is not trust. In that respect one familiar definition seems insufficiently circumscribed: 'When we say we trust someone or that someone is trustworthy, we implicitly mean that the probability that he will perform an action that is beneficial or at least not detrimental to us is high enough for us to consider engaging in some form of cooperation with him' (Gambetta 1988: 217). It is possible to believe that people will act as we wish, but if they do so for purely prudential or instrumental reasons (for example because they would otherwise be punished, or hope to obtain a reward), complete confidence is consistent with an absence of trust.

7.16 To recap, trust is a belief that other people hold pro-social attitudes that we can rely on. The empirical analysis confirms that there is a clear association between the perceived fairness of others and generalized trust. But it is evident that trust is a function of average as well as individual perceptions, and so depends on the national context. The importance of culture is consistent with the idea that trust is rooted in a moral evaluation of people in general, and so often qualifies as an attitude rather than being just a belief about the attitudes of others.

Conclusion

8.1 The sociology of attitudes is both well established and surprisingly underdeveloped. Social scientists have done research on attitudes without worrying too much about what they are, which is no bad thing. The time seems right, however, to reflect on the concept. What attitudes are affects how they should be measured and how they - and in turn behaviour - can be changed.

8.2 By contrast with the individual likes and dislikes described by psychologists, the characterization offered here is highly social. In holding attitudes, we think about society and the rights and duties of others as well as ourselves. The judgements may be personal, but they are formed in and relate to a social context. They are neither objective matters of fact nor subjective matters of taste; they express values. Attitudes emerge from the interaction of beliefs, preferences, behaviour and values at the individual level, but these influences are themselves formed through the interaction of culture, human nature and the world around us.

8.3 Understanding the nature of social attitudes helps in deciding what to measure and how. In some instances it may be important to collect data on beliefs or preferences; in others we may wish to ensure that attitudes are clearly distinguished from them. The direct link between latent values and observed attitudes gives us an additional motivation for investigating the genesis of attitudes; sociologists have so far made only modest contributions to the study of values (which has been dominated by psychologists and political scientists, e.g. Milton Rokeach, Shalom Schwartz, and Jonathan Haidt on the one hand and Ronald Inglehart, Pippa Norris and Christian Welzel on the other).

8.4 This account also helps to clarify the status of other key concepts, including tolerance and trust. Tolerance is an attitude, but generalised intolerance more closely resembles a feeling about others. Conversely trust is a belief about attitudes, but generalised trust is as much a moral assessment of humanity (i.e. an attitude) as it is a belief about how other people will behave. A great deal remains to be done, but the sociology of attitudes has a compelling future.

References

AJZEN, I. & Fishbein, M. (2005) 'The influence of attitudes on behavior', in Albarracin, D., Johnson, B.T. & Zanna, M.P. (Eds.) The Handbook of Attitudes, Mahwah, NJ: Lawrence Erlbaum Associates.

BACHARACH, M. & Gambetta, D. (2001) 'Trust in signs', in Cook, K.S. (Ed.) Trust and Society, New York, NY: Russell Sage Foundation, p. 148 - 184.

BASSILI, J.N. & Brown, R.D. (2005) 'Implicit and explicit attitudes: Research, challenges, and theory', in Albarracin, D., Johnson, B.T. & Zanna, M.P. (Eds.) The Handbook of Attitudes, Mahwah, NJ: Lawrence Erlbaum Associates.

DEUTSCHER, I. (1973) What We Say/What We Do: Sentiments & Acts, Glenview, IL: Scott, Foresman and Company.

EAGLY, A.H. & Chaiken, S. (1993) The Psychology of Attitudes, Fort Worth, TX: Harcourt Brace Jovanovich.

EAGLY, A.H. & Chaiken, S. (2005) 'Attitude research in the 21st century: The current state of knowledge', in Albarracin, D., Johnson, B.T. & Zanna, M.P. (Eds.) The Handbook of Attitudes, Mahwah, NJ: Lawrence Erlbaum Associates.

GAMBETTA. D. (1988) 'Can we trust trust?', in Gambetta, D. (Ed.) Trust: Making and Breaking Cooperative Relations, Oxford: Basil Blackwell.

GRAHAM, J., Haidt, J., Koleva, S., Motyl, M., Iyer, R., Wojcik, S., & Ditto, P. H. (2013) 'Moral foundations theory: The pragmatic validity of moral pluralism', Advances in Experimental Social Psychology, Vol. 47, p. 55 - 130. [doi://dx.doi.org/10.1016/B978-0-12-407236-7.00002-4]

INGLEHART, R. & Welzel, C. (2005) Modernization, Cultural Change and Democracy: The Human Development Sequence, Cambridge: Cambridge University Press.

KROSNICK, J.A, Judd, C.M. & Wittenbrink, B. (2005) 'The measurement of attitudes', in Albarracin, D., Johnson, B.T. & Zanna, M.P. (Eds.) The Handbook of Attitudes, Mahwah, NJ: Lawrence Erlbaum Associates.

KRUGLANSKI, A.W. & Stroebe, W. (2005) 'The influence of beliefs and goals on attitudes: Issues of structure, function, and dynamics', in Albarracin, D., Johnson, B.T. & Zanna, M.P. (Eds.) The Handbook of Attitudes, Mahwah, NJ: Lawrence Erlbaum Associates.

LAPIERE, R.T. (1934) Attitudes vs. actions, Social Forces, Vol. 13, No. 2, p. 230 - 37. [doi://dx.doi.org/10.2307/2570339]

MARSH, K.L. & Wallace, H.M. (2005) 'The influence of attitudes on beliefs: Formation and change', in Albarracin, D., Johnson, B.T. & Zanna, M.P. (Eds.) The Handbook of Attitudes, Mahwah, NJ: Lawrence Erlbaum Associates.

MILLER, A.S. & Mitamura, T. (2003) 'Are surveys on trust trustworthy?', Social Psychology Quarterly, Vol. 66, No. 1, p. 62 - 70. [doi://dx.doi.org/10.2307/3090141]

PROCTER, M. (2008) 'Measuring attitudes', in Gilbert, N. (Ed.), Researching Social Life, Third Edition, London: Sage.

RYLE, G. (2000[1949]) The Concept of Mind, London: Penguin Classics.

SCHWARTZ, S.H. (2007) 'Value orientations: Measurement, antecedents and consequences across nations', in Jowell, R., Roberts, C., Fitzgerald, R. & Eva, G. (Eds.), Measuring Attitudes Cross-nationally: Lessons from the European Social Survey, London: Sage. [doi://dx.doi.org/10.4135/9781849209458.n9]

SCHWARTZ, S. H. (2012). 'An overview of the Schwartz Theory of Basic Values', Online Readings in Psychology and Culture, Vol. 2(1). http://dx.doi.org/10.9707/2307-0919.1116 [doi://dx.doi.org/10.9707/2307-0919.1116]

USLANER, E.M. (2002) The Moral Foundations of Trust, Cambridge: Cambridge University Press.

USLANER, E.M. (2008) 'Trust as a moral value', in Castiglione, D., van Deth, J.W. & Wolleb, G. (Eds.), Handbook of Social Capital, Oxford: Oxford University Press.

WILLIAMS, B. (1985) Ethics and the Limits of Philosophy, London: Collins / Fontana Press.

YAMAGISHI, T., Kikuchi, M. & Kosugi, M. (1999) 'Trust, gullibility, and social intelligence', Asian Journal of Social Psychology, Vol. 2, p. 145 - 61. [doi://dx.doi.org/10.1111/1467-839X.00030]

© Sociological Research Online, 1996-2014

UniS: University of Surrey logo University of Stirling logo British Sociological Association logo Sage Publications logo Electronic Libraries Programme logo Epress logo