Talent, Effort or Social Background?: An empirical assessment of popular explanations for educational outcomes Open Access

Despite persistent social inequalities in educational attainment in most Western societies (Pfeffer 2008), scholars have described, in a variety of terms, how education as an institution contributes to a representation of society in which success is ‘achieved’ on meritocratic grounds (e.g., Bourdieu 1984; McNamee and Miller 2009). This literature points among other things to the way neoliberal elements (the stress on individual choice, self-reliance, hard work, etc.) slip into educational practice and conceal the often strong social inequalities in educational outcomes or even contribute to their persistence (e.g., Bradbury 2013; Boone and Van Houtte 2013). In this way, the ongoing ‘massification’ of education is thought to fundamentally alter the perception of social inequalities in modern societies by pushing citizens to rely on more individualistic interpretations of success and failure. This tendency could be highly consequential for: (1) the perceived legitimacy of and public support for policies aimed at combating social inequalities in education (Appelbaum 2001), (2) the interaction between education-based groups (e.g., Stubager 2009; Tannock 2008) and (c) the individuals’ striving for an improvement in personal status (Kluegel and Smith 1986). However, surprisingly, little research has been conducted into popular explanations for educational outcomes among the public at large (but see Bourdieu 1984: 386–98).

This paper aims to partially fill this gap by studying popular support of three explanations for educational outcomes (i.e., success and failure): talent, effort and social background. Survey research was used to assess the support for each of them in Flanders (the Dutch-speaking part of Belgium), how these explanations related empirically to one another and the existing social differences in support.

Popular explanations focus on the ‘why’ of things (Kluegel and Smith 1986). They are, therefore, an important constituent element of citizen's stratification beliefs (McWha and Carr 2009; Niemelä 2008) and are often studied for phenomena such as poverty, wealth, unemployment or neediness (e.g., Bullock 1999; Kluegel and Smith 1986; Kreidl 2000). To date, little research has been conducted into public opinion on explanations for educational outcomes (but see Bourdieu 1984; Thrupp et al.2003). This paper attempts to partially fill this gap. We do not intend to draw up an exhaustive list of explanations for educational outcomes, but focus instead on support for three elements: talent, effort and social background. That choice is informed by the insight that public opinion is to some extent influenced by public discourse surrounding a specific topic (Zaller 1992). Consequently, we should focus on explanations for which it may be shown that they are in some way supported by (1) the way education is talked about by public actors like the media, political parties and educational policies and (2) institutionalised educational practices.

With respect to the latter, an extended literature has documented the subtle ways in which educational practice allegedly institutionalises highly individualistic conceptions of the self, thereby suggesting that talent and effort are the key ingredients of educational success (e.g., Bradbury 2013; Meyer 1977). Bourdieu (2000), for example, points to (1) the repeated and detailed testing of educational outcomes throughout the educational career, (2) the increasing use of intelligence tests which attempt to measure children's scholastic potential and by doing so ‘naturalise’ educational outcomes (Bourdieu 1990), (3) the content of education with its emphasis on creativity (Bourdieu 1984: 415), (4) an individual-centred approach with strong emphasis on individual choices and (5) ‘limited social mobility’ which conceals the often strong intergenerational reproduction in educational outcomes (Bourdieu 1998: 21–2). These characteristics combine in a ‘culture of the gift’ wherein educational outcomes are presented as the result of superior and natural qualities (Reed-Danahay 2005: 48). In this way, education is thought to inculcate explanations for educational outcomes favouring strictly individualistic properties, e.g., effort and talent, and concealing more structural explanations in terms of the relevance of pupils’ social background and its mismatch with conventional educational practice.

The relevance of social background to educational outcomes, however, is not completely absent in public discourse. Researchers studying inequality in education have expressed the equity and equality of opportunity in education in terms of the predictability of educational outcomes on the basis of the social background characteristics of students (e.g., Boone and Van Houtte 2013; Pfeffer 2008). In Flanders, the region we have studied, one of the key objectives of current plans to reform secondary education is aimed at weakening the strong association between children's social background and their educational attainment (Smet 2010). The feasibility, priority and, particularly, the amount of investments that should be given to this purpose became the object of intense public debate. The extent to which these debates increased awareness for social reproduction in educational attainment among the public at large is an empirical question. The preceding arguments do, however, suggest that all three considered explanations are in some way present in public discourse rendering it likely that citizens are familiar with them. Let us now consider these three explanations from a more analytic point of view.

Although discussion remains about how they exactly combine, variations in popular explanations are thought to be structured by at least two fundamental dimensions. The first refers to who is held responsible (individuals–society), the second to the degree in which those responsible have a grip on the situation (blame–fate; van Oorschot and Halman, 2000). Explanations for educational outcomes in terms of talent and effort are both clearly individualistic. They differ, however, in the extent to which the individual is thought to have control over the situation. Attributing educational outcomes to effort, and a lack of it, assumes that the situation is hardly beyond the control of the individual. When it is used to explain failure, it contains a clear blame component. In contrast, talents are seen as stable innate qualities which can be trained and developed, but whose presence is beyond the control of the individual. This renders an explanation in terms of talent much more fatalistic. We chose to examine the people's views of the role of talent rather than intelligence or ability, since the term ‘intelligence’ is being avoided in current policy discourse in Flanders. Policy-makers primarily refer to talents, and in doing so, they are stressing that people do not have more or less but different talents (Smet 2010).

A social background explanation for educational differences clearly articulates a fatalistic element. Whether the individual or the system is to blame, however, is more difficult to determine. Indeed, in public discourse, a social background explanation appears in two variants. In the first, the insights gained from the literature on cultural capital are used to argue that where educational inequality is concerned, education is fighting a battle it cannot win. Such fatalistic views were also found among teachers when asked to what extent they considered themselves accountable for educational outcomes (Thrupp et al. 2003). Referring to the social background of students was another way of saying ‘this is beyond education's control’. In this interpretation, the background of pupils is considered an individual resource for which, to borrow Basil Bernstein famous quote, ‘education cannot compensate’. In public debate, that argument is often countered by referring to the fact that countries do differ in the extent to which pupils’ social background predicts educational outcomes. Even if accomplishing real equality of opportunity in education is to remain a utopian ideal, so the second interpretation of a social background explanation goes, a lot of potential for improvement remains and the educational system should – at least partially – be held responsible.

In this paper, we are not able to discriminate between both variants and therefore mainly focus on the ‘non-innate fatalistic’ element in a social background explanation for educational outcomes. This implies that a purely structural explanation, in terms of van Oorschot and Halman's (2000) typology, is not present in our analysis. In the next section, we discuss expectations regarding social differences in the support each of the distinguished explanations received.

Popular explanations are assumed to be part of the collective representations through which people perceive themselves and their society (Feather 1983: 229): ‘blaming the person is not a matter of erroneous reasoning but a matter of drawing a logical conclusion from the commonly accepted system of representations’ (Moscovici 1982: 133). These representations are thought to be supplied by societal institutions such as the media, welfare state, political parties and education (Hewstone and Augoustinos 1998: 63). In that context, conflict sociologists maintain that the dominant ideas in society generally support the societal status quo by promoting individual explanations (Jackman 1994). To dominant groups, individual explanations have two appealing characteristics: (1) they essentialise differences and reduce them to individual characteristics of a person, so that the differences are regarded as justified, and (2) precisely because of that, they instigate interpersonal competition. Both elements undermine group formation and collective action. This might especially apply to educational differences for the reasons mentioned above. Bourdieu (1984: 291), for example, considers educational credentials a form of ‘objectified’ symbolic capital: ‘[…] a reputation for competence and image of respectability and honourability’. Symbolic capital provides dominant groups with social status; it converts power into prestige by depicting realisations as an outcome of ‘natural gifts’ (Bourdieu 1990). As the product of acts of mutual recognition, symbolic capital is by definition unstable. The distinguishing feature of educational credentials when compared to other forms of symbolic capital, then, is that they are ‘objectified’ (Bourdieu 1998: 50–1): ‘As an official definition of an official identity, it [a school diploma] frees its holder from the symbolic struggle of all against all by imposing the universally approved perspective’ (Bourdieu 1989: 21–2). Consequently, allocation based on educational credentials is expected to be considered legitimate by both the higher and less educated.

The preceding arguments conjure up two expectations: (1) explanations in terms of talent and effort are more strongly supported than the social background explanation (Hypothesis 1) and (2) the differences between social groups in support of each of the distinguished explanations are relatively small (Hypothesis 2). Several authors have emphasised that one of the characteristics of popular explanations is the modest amount of explained variance by social-positional characteristics, irrespective of the specific phenomenon (e.g., Niemelä 2008; Shirazi and Biel 2005).

The last expectation does not, of course, rule out that such differences do actually exist. Micro explanations focus on the differences between groups within the same society. They differ in the extent to which causal attributions are thought to reflect self-interest or socialisation effects. Below, we discuss both strands separately and show that they lead to contradicting hypotheses regarding to the expected social differences in support of the three explanations for educational outcomes.

When popular explanations are thought to primarily reflect self-interest, successful people are expected to adhere more to explanations that represent success in terms of ‘natural’ or superior individual qualities (e.g., Bullock 1999). Following the same reasoning, vulnerable people stand to gain from explanations that attribute responsibility to external influences (e.g., Furnham 1982). Since more structural and individual explanations usually do not strongly correlate negatively, it is often thought that the stronger support of structural explanations among socially weaker groups will not result in those groups renouncing individual explanations (Kluegel and Smith 1986: 93). Self-interest does not solely have to relate to the objective social position. The relationship between popular explanations and class identification (e.g., Nasser 2007), or the impact of ideological egalitarianism (Lee et al.1990), is often interpreted in the same way. Such ideological preferences have been shown to play a role in the evaluation of more structural explanations (e.g., Kreidl 2000).

To the extent that people adhere to popular explanations that match their own social position, it can be expected that, for the explanation of educational differences, highly educated people and those who feel attached to them will refer more to explanations in terms of talent and effort (Hypothesis 3a). These explanations grant them the most social status. Indeed, whether an explanation in terms of social background attributes success to characteristics which fall beyond control of the individual (as contrasted with effort) or beyond the physical inborn qualities of the individual (as contrasted with talent), both interpretations undermine the ‘deservingness’ component of being higher educated. Moreover, as explained earlier, at least one strand in public discourse formulates a structural explanation for educational outcomes in terms of the continuing relevance of social background for educational success. Therefore, we expect the higher educated to renounce them (Hypothesis 3b). The opposite is expected for less educated people or those who feel an affinity with them. In addition, it can be expected that people who believe that our society should become ‘more equal’, and that it is the government's task to realise that, will put a stronger emphasis on social background and (to a lesser degree) renounce talent and effort explanations (Hypothesis 4a and b).

Aside from a well-understood self-interest, however, popular explanations may also be partly regarded as social knowledge or insight. In Flanders, as in most West European regions, the social background of youngsters is quite strongly related to educational outcomes (Pfeffer 2008). Seen from this point of view, acknowledging that social background is relevant for educational outcomes is a form of social knowledge. The completion of higher education, in particular, is expected to increase social knowledge and insight (for an overview, see Pascarella and Terenzini 2005). Finding a net relationship between education and a more structural explanation is often interpreted in this way (Kluegel and Smith 1986): discovering ‘schooling as a conservative force’ presupposes some experience of ‘schooling as a liberating force’ (Bourdieu 1984: 390). This reasoning leads to the expectation of finding more support for a social background explanation for educational outcomes among the higher educated (Hypothesis 5).

Social knowledge and insight, however, do not develop equally in all fields of study. A number of researchers reported systematic differences in the popular explanations adhered to according to the field of study (McWha and Carr 2009). People who completed more market-oriented studies, as opposed to social studies, have a stronger tendency to refer to individual explanations (Guimond and Palmer 1990; Lopez et al.1998; Seider et al.2010). Therefore, we expect that people who studied humanities rather than economics or law will more strongly endorse explanations for educational outcomes in terms of social background (Hypothesis 6a) and renounce explanations in terms of talent and effort (Hypothesis 6b), irrespective of their actual level of educational attainment.

As regards to the differences in educational groups, the foregoing leads to hypotheses that contradict those that assume popular explanations to be a function of self-interest. Nevertheless, those explanations do not necessarily exclude one another, precisely because we know that the different popular explanations are often not each other's alternative. Research on popular explanations for wealth and poverty found that higher education reduces the support for an individual, as well as a structural explanation, as predicted by the socialisation explanation and the self-interest thesis, respectively (Hunt 2004).

Education in Flanders is organised centrally and free of charge during compulsory education (aged 6–18). Truly elitist private schools hardly exist in Flanders – i.e., only 0.14% of the pupils in secondary education is enrolled in a non-subsidised private school (De Ro 2008) –and tuition fees for higher education are less than half of the EU21 average (OECD 2013: 144). Those features might render the role played by social background for educational outcomes less visible to people. ‘Less visible’ is applicable here, as comparative research has shown that educational mobility in Belgium (and, hence in Flanders) is one of the lowest in Europe (Pfeffer 2008) with mobility also hardly increasing over time (Groenez 2010). Social inequality in educational outcomes in Belgium primarily results from the early and strong tracking in secondary education (first choice at age 12). Through a long series of selections, small social differences in achievements and educational choices accumulate (see Boone and Van Houtte 2013). In Flanders, every person who obtains a diploma secondary education is allowed to enrol in all types of tertiary higher education. In practice, however, the enrolment rates in higher education differ sharply according to the educational track one followed in secondary education. They are above 90% for pupils who followed the general track, much lower for technical education (about 60%) and lowest for pupils who followed vocational education (about 30%). From the third grade in secondary education onward (aged 14), it remains possible for pupils to ‘descend’ from general to technical and further to vocational education; ‘ascending’ movements in the opposite direction almost never occur.

The absence of strong financial and structural barriers and the strong social reproduction of educational outcomes renders Flanders an interesting case study, as they maximise the probability that educational outcomes will not simply be seen as resulting from material class differentials. In fact, were education to take root anywhere as ‘objectified symbolic capital’, then Flanders appears to possess all the prerequisites.

In order to test our hypotheses, we rely on data from a survey of a simple random sample of the population of Flanders aged 18–75 gathered in the fall of 2010 (for technical report, see Claeys et al. 2011). The sample was drawn from the National Person Register, which contains records of all Belgian citizens. A mail questionnaire was used and 1967 properly filled out questionnaires were returned (net response rate 49.4%). Subsequent comparisons with population data revealed that younger people, males and the lower educated were somewhat under-represented in our sample. Therefore, weights based on the combination of age, gender and the educational level are used (36 coefficients were calculated; 5 of them exceed 2, and the highest equals 3). In order to work with a homogenous sample in terms of actual educational experience, we excluded students (n = 70) from the analysis.

With respect to the educational position of respondents, we use the highest level attained (six categories, with ‘missings’ as a separate category) and the field of study for people who followed secondary education (nine categories). The latter is a slightly modified version of the one used in the European Social Survey. The economic position is measured by a composed indicator based on the average monthly disposable family income, house ownership, monthly savings and the last occupation (EGP classification). A categorical principal component analysis on these characteristics revealed one dimension that captured 55% of the variance (Cronbach's alpha = 0.729). Respondents’ educational background was measured by means of the educational level of their mother.

Egalitarianism is measured with five items (e.g., the government should reduce income differences) which are commonly used in public opinion research (Cronbach's alpha = 0.784). A high score on this scale reflects a traditional left-wing ideology and expresses a longing for equality also with the idea that it is the government's task to accomplish this. With respect to the treatment of missing values for the attitude scales, we compared the results obtained using list-wise deletion where mean value imputation was used. Since differences turned out to be very small, we decided to use list-wise deletion.

Finally, an indicator for education-based group identity, developed by Stubager (2009), was included in the regression models. People were asked whether they felt attached to people with high or low education. In Flanders, this indicator yields a lot of ‘missings’, ‘don't knows’ or ‘do not feel close to these educational groups’. We included them as separate categories in the regression models. All attitude scales were entered as factor scores in the regression analysis.

In order to measure the three explanations discerned, we asked people to rate different statements referring to the three theoretically discerned explanations on a scale from 1 (totally disagree) to 5 (totally agree). All items refer to how the situation is rather than how it should be. When formulating the items, it was ensured that both success and failure were addressed. The latter is informed by the observation that explanations for success and failure sometimes differ (e.g., Evans 1997; Kreidl 2000). Success is attributed to individual factors more often than failure (e.g., Hunt 2004), and this also applies to educational differences. Table 1 presents the frequency distribution for all items.

Two items (in italics) are more general and do not clearly articulate one of the theoretically discerned explanations. They have in common that they create an image of a situation (equal educational opportunities versus a lack of control over educational success) without precisely explaining how that situation arises. They were included because of the strong notion supported in the literature that socially weaker groups often have a diffuse feeling of ‘fatalism’ (Charlesworth 1999), but were excluded from the scale construction.

In general, people in Flanders stressed strongly the importance of effort and dedication for educational success. Nearly 80% of the people in Flanders believed that effort and dedication are the most important factors for obtaining a diploma (Accept H1). The ‘most important factors’, however, clearly cannot be equated with ‘the only factors’. A mere 43% believed, that in principle, everyone is able to obtain a diploma, and only 13% were convinced that not doing so was due to laziness or a lack of effort. Items that referred to an actual inequality in educational opportunities and the role of social background in obtaining a diploma were widely approved partly because of this. Just under half of the respondents confirmed that children from wealthy parents had a greater chance of success in school, and more than 40% also acknowledged the importance of parents’ educational level. Given these percentages, it seems that in explaining educational differences, the average person in Flanders does not see any contrast between an individual explanation and an explanation that refers to the role of social background.

People in Flanders held a somewhat ambivalent attitude towards the role of talent in the explanation of educational differences. Talent was referred to when explaining success, and a substantial group also acknowledged that talent plays a role in success and that, therefore, some people will never succeed. Only 6%, however, agreed to the idea that talent is the decisive factor for failure or success.

The use of ‘mainly’ and ‘most’ in the wording of the items did not prevent the respondents from answering as if the explanations were not competing alternatives. This tendency has also been observed in other research (Kluegel 1990).

Subsequently, an exploratory factor analysis on the 11 items revealed three factors with an eigenvalue above 1 with the items belonging to one of the three putative dimensions (talent, effort and social background) also loading on the same factor. The use of an oblique rotation allowed us to study the relations between the dimensions found. Only the one between talent and social background was significant and substantial (r = 0.387). Considering the correlation structure – a strong correlation between talent and background, and virtually no relationship with an explanation in terms of effort – people in Flanders did not seem to make a distinction between ‘meritocratic’ – i.e., talent and effort – and ‘non-meritocratic’ – i.e., social background – elements but rather between characteristics that are ‘given’ – which are beyond the control of the individual, i.e., talent and social background – and factors which can be manipulated by individuals, i.e., effort. In doing so, the degree to which they attached importance to ‘given’ characteristics said practically nothing about the role they attached to ‘controllable’ characteristics.

Regression analyses were used to assess the social anchorage of the three explanations. As expected, the explained variance proved to be low, indicating the general picture that appeared from the frequency distribution was strongly underpinned by all social groups (Accept H2). Moreover, the variance explained was the lowest for the effort-related explanation – the explanation that has been the most widely supported in public opinion (Table 2).

The first model only included gender, age and the educational level. Gender acted as a control variable. With respect to age, it has been repeatedly found that older people highly supported individual explanations for both success and failure (e.g., Cozzarelli et al. 2001; Kreidl 2000). The same applied to the explanation for educational differences. Older groups reverted to explanations in terms of talent far more often but, at the same time, also adhered more to explanations that emphasised the role of social background. Older cohorts thus seem to put more emphasis on ‘given’ factors – aspects that are important, but cannot be controlled by the individual. When we included in each model the other two discerned explanations as control variables, and thus gauged the true uniqueness of each explanation, the effect of age on the background-related explanation became insignificant, while the association between age and the talent-related explanation barely weakened. Although our data do not allow for an empirical test, there are reasons to believe this might concern a cohort effect. Older cohorts attended school in a period in which mass education had just started to blossom. A considerable number of people in these cohorts did not go to school for very long, because the family situation simply did not allow it. Over the past decades, these people have witnessed numerous initiatives to improve educational opportunities and combat inequality in education. Perhaps for that reason they tend to believe that, today, school success is primarily determined by talent.

Education was related to every explanatory variable, but for the explanation in terms of talent and social background alone, that relationship could be interpreted unambiguously. The higher educated were less inclined to explain school success or failure on the basis of talent, and noticeably more often reverted to an explanation in terms of social background. For the latter, the correlation was non-linear with a clear gap between those who completed higher education and those who did not. That pattern was also predicted by the socialisation perspective (Accept H5) and contradicts the explanation in terms of self-interest (Reject H3a and b). This also implies that in our society a talent-related explanation for educational outcomes is supported most by the less educated. This leads to the odd proposition, but one that was also found by others, that the legitimisation of the position of the highly educated is actually primarily realised by the less educated (Bourdieu 2000: 170–5).

The other relevant indicators were added in a second model. The field of study seemed to be important for all explanations, and in terms of talent, it even explained away the effect of the educational level. The direction of the correlations was as expected on the basis of the literature. Those respondents who studied social sciences, education or arts were inclined to revert more to an explanation in terms of social background and less to an explanation in terms of talent or effort (Accept H6a and b). The reverse applied to respondents who completed study programmes in, or related to, economics, sciences and applied sciences, the differences between the correlation directions for both groups being the greatest for the dimension of effort. Interestingly, within the group of humanities (education, social sciences and arts), the greatest differences existed for the effort-related explanation – which is consistent with authors who claim that the renouncing of an individual statement is also related to an affinity with statistical methods and particularly with the notion that individual variance can show a clear, although never conclusive pattern (e.g., Guimond and Palmer 1990). This type of knowledge is conveyed far more in the humanities than in studies related to education or arts.

The economic position was not significantly related to any of the explanations; adding it to the model did not influence the correlations between educational characteristics or the three explanations. The same applied to the educational background of people – i.e., the educational level of their parents. We found a small, yet virtually negligible effect on the background-related explanation in those cases where the mother had a higher education diploma. Respondents whose mother followed higher education were slightly more inclined to acknowledge the role of social background. We also explored interaction effects between the educational background and own educational position (i.e., the effect of social mobility), but these effects turned out insignificant.

Ideological egalitarianism did correlate significantly and logically with the explanations for educational differences. People who strive for more equality emphasised the background dimension in particular and, at a later stage, were less inclined to attribute school success exclusively to the role of effort, although the latter was clearly less strongly ideologically determined than the first (Accept H4a and b).

Attachment to education-based groups solely correlated with the explanation in terms of background. Those respondents who felt a strong connection with the highly educated supported this explanation more than people who did not feel any connection with one of the educational groups. The difference between the groups who felt a connection with respectively the highly and the less educated was not significant. We did not find an interaction between the educational level and the degree of educational connection either.

This paper assessed popular support for three different explanations for educational outcomes, namely talent, effort and social background. Our research attempted to answer three different questions. The first and the second pertain to the overall distribution of the distinguished explanations and their mutual relationship. They were informed by scholar's warning that despite persistent social differences in educational outcomes, the ‘massification’ of education itself possibly transforms the perception of social inequality in general and inequalities in educational outcomes in particular.

On average, three quarters of the people in Flanders referred to effort and dedication as the main factors for educational success. The Flemish emphasised in particular that effort and dedication are necessary conditions for success. They added that these are not sufficient conditions. Only a minority attributed educational failure exclusively to laziness, and the role of social background was also amply recognised. Talent was deemed relevant, but definitely put into perspective. Only a small percentage endorsed the notion that educational failure is due to a lack of talent. Therefore, although the well-documented emphasis on meritocratic elements in public discourse on educational success takes root in the way people explain success, this does not necessarily imply the adoption of a harsh perspective with respect to educational failure. This suggests that justifying a policy for equal educational opportunities by referring to the ineradicable role played by social background, in light of educational failure, bears the potential to find considerable approval among the public at large.

In this research, we did not attempt to draw up an exhaustive list of explanations that people give for educational differences, but focused on three important explanations. Further research should search for other popular explanations. Their existence is suggested by the finding that the correlation between the explanations that refer to talent and background, on the one hand, and effort, on the other hand, is not positive or negative, but simply non-existent. In the way in which people reconcile these explanations, a wide range of ‘interim explanations’ probably lies hidden that will comprise elements of both types of explanation (Lepianka et al.2009). One of the limitations of this paper, in this context, is that we lacked a measure that tapped into a purely social/structural explanation concerning the role of the educational system in ensuring equality of opportunity. As explained earlier, the main reason for this is that in Flanders there are no large financial barriers towards education, there are almost no pure elite schools and realising equality of opportunity is, already for a long period of time, very high on the agenda of educational policy. Hence, if today's educational system is represented as unfair in public discourse, it is almost always in terms of education's failure to cope with differences in cultural and economic resources of pupils (i.e., social background). In this way, a purely structural explanation is lacking in public discourse. It is, however, perfectly possible that despite all these the educational system is perceived to be structurally unfair by a certain part of the population. Therefore, further research should include a measure to tap into these feelings, especially when the regions studied are characterised by high tuition fees or large differences between public and private schools (e.g., the UK).

The third question of this paper concerned whether and how dominant groups, the higher educated, used popular explanations for success to justify their own position. We did not find strong support for the idea that the dominant groups ‘rationalise’ their position by referring to individual and even natural features as the determinants of their success. On the contrary, like Bourdieu (1984), we found that the belief that social background is still important for educational success is mainly supported by those who have been most successful in education. At first sight, these findings contradict the more general literature on popular explanations for phenomena such as neediness, wealth or poverty. Our findings, however, becomes less surprising when it is interpreted through the lens provided by literature on social domination and reproduction.

Dominant groups, according to Jackman (1994), are caught between the need to differentiate themselves sufficiently to procure inequality and the need to integrate the dominated within a cooperative relationship in order to preserve the flow of benefits. Thus, dominant groups tend to develop attitudes that can best be described as ‘paternalism’, a conditional appreciation in which the dominant group hangs on to a positive picture of the dominated group for as long as the dominated group does not fundamentally question the dominant group's socio-political position. This is exactly what we found. The higher educated do not have to take a harsh position in this matter; the less educated and institutional practice do ‘the dirty work in reproducing privilege and disadvantage’ (cf. DiMaggio 2012: 12) for them. Indeed, Bourdieu (1984: 65–6) articulated his pessimism about education's ultimate capacity to curtail social inequality because the attitudes, dispositions and tastes acquired in education never compensate for the disadvantages associated with a culturally deprived social background. It was precisely that what educational policies have attempted to do: to remedy or compensate for the alleged deficiencies among working and lower-class pupils. The theoretical fairness of a meritocratic-organised society has not only been the motivating force behind the push to expand access to education for all. It also led to attempts to mitigate the influence of non-prescriptive criteria by providing additional support within education, which in practice implied a broadened scope for education (cf. Pring 2010). This ‘educationalisation of society’ (cf. Smeyers and DePaepe 2008) continues to confirm and justify education's centrality and status in general, and that of the higher educated in particular, in contemporary societies. It is for this reason that the higher educated are able to criticise education without this implying a reduction of their own status. It is precisely the attitude adopted by the higher educated such as we found, Jackman (1994: 180) would say, that strongly restricts change as (1) the dominated are not obliged to distance themselves to acquire a positive image of their own group, as the dominant group does not feed into the negative image of the dominated and (2) the dominated themselves have to take the initiative to induce conflict in the relationship. Under these circumstances, ‘education-based discrimination’ (Tannock 2008) is likely to prevail.

At the more general level, the foregoing interpretation also contributes to the general literature on popular explanations by underscoring the importance of adopting a more relational approach when interpreting social differences in support for different explanations. Indeed, only when one takes into account the explanations the lower educated adhered to, a full understanding of the position of the higher educated is possible. Differences in support for popular explanations with respect to socialisation or self-interest cannot be considered separately but are closely intertwined. Follow-up research should explore the merits of this interpretation by comparing social differences in support of explanations for educational outcomes with those containing phenomena which are more often studied.

Bram Spruyt is an assistant professor of Sociology at the Vrije Universiteit Brussel (Belgium) and member of the research group TOR. His main research interests include the sociology of education, cultural sociology and public opinion research. This paper is the outcome of a research project on education-based group identity. This project aims to assess how education-based social categories are used in daily interactions and contribute to the social reproduction of educational differentials in among other things political behavior, social participation and opinions.