Contextual effects on subjective social position: Evidence from European countries

Abstract

This study investigated subjective social position in 21 European countries using data from the social inequality module of the International Social Survey Programme 2009. Subjective social position shows people’s self-location in a social hierarchy. Most studies on subjective social position have typically involved a few countries and neglected the role of national educational and occupational structures. We hypothesise that these characteristics, together with national-level economic factors, modify the effect of individual-level characteristics on subjective social location. Our expectations are based on extended reference group thesis, the big-fish-little-pond and stigmatisation arguments. The results of multi-level analysis indicated that the extent of educational and income inequalities in society as well as occupational structures influence the importance of individual incomes, education and occupations for estimating social position.

Keywords

Country-level characteristics European countries extended reference group multi-level analysis social comparison social hierarchy subjective social position

Introduction

In recent years, the question of rising inequalities has received a lot of attention in both scholarly discussions and wider media public debates. The OECD (2011) claims that in the 2000s the gap between rich and poor widened even in some traditionally low-inequality countries in continental and Northern Europe. In the light of rising inequalities, it is informative to look beyond traditional measures of inequality and explore how people themselves estimate their position in societies with varying educational and labour market contexts.

Subjective social position indicates how people perceive their relative position in the social hierarchy. In order to evaluate their social position, people make comparisons in societal contexts. The concept subjective social position is often used in studies of social and political attitudes as well as in studies of subjective health (Centers, 1953; Davis and Robinson, 1988; Evans et al., 1992; Hodge and Treiman, 1968; Hout, 2008; Jackman and Jackman, 1983; Vanneman and Cannon, 1987). Political behaviour depends more on how a person feels socially than on his or her position according to objective characteristics such as education, occupation or income (Matějů and Večernik, 1999). Subjective social position influences aspects such as self-related health (Goodman et al., 2007; Kawachi et al., 1999; Operario et al., 2004). Perceived social position is associated with both psychological and physiological functioning, because low subjective standing is associated with greater stress (e.g. Adler et al., 2000). So, an important question is what factors affect how people evaluate their position in society.

Early attempts to map out the predictors of subjective social position studied individual-level characteristics and relied, for the most part, on single-country analyses (Davis and Robinson, 1998; Goldman et al., 2006; Hout, 2008; Kikkawa, 2000; Knudsen, 1988). In addition to standard measures of objective status, such as education, occupation and income of respondents, studies examined impact of other socio-economic variables (such as union membership, home ownership, race, ethnicity, age, gender etc.), as well as measures of the respondents’ spouses status (see Abbott, 1987; Baxter, 1994; Beeghley and Cochran, 1988; Centers, 1949; Davis and Robinson, 1998; Hodge and Treiman, 1968; Yamaguchi and Wang, 2002). Findings suggest that socio-economic variables are typically significant predictors of subjective social position but the explanatory power of these variables is quite modest and has decreased (see Goyder, 1975; Jackman and Jackman, 1983). Researchers explain this pattern by the reference group processes such as individualisation (Oddsson, 2010). However, despite the burgeoning empirical evidence, it is not entirely clear whether the predictors of subjective social position are similar across population groups and social contexts.

Earlier studies mostly drew from reference group theory assuming that people make comparisons on the basis of family, colleagues and friends rather than with society as a whole or abstract groupings within it (see, for example, Evans et al., 1992). However, more recently, several authors suggested that social developments such as globalisation and the rise of the information society effects a range of comparisons people are able to make (Delhey and Kohler, 2006; Suls et al., 2002; Sweeney and McFarlin, 2004). This means that social comparisons are situated in a broader social context that extends well beyond the immediate social environment, indicating the importance of cross-national analyses (Buunk and Mussweiler, 2001; Fahey, 2007).

Some previous research shows that besides individual objective social position, economic and social context in society is important. Evans and Kelley (2004) and Andersen and Curtis (2012) studied the effect of some indicators of national economic context, such as gross domestic product (GDP) per capita, income inequality and national level of unemployment. They find that all of these indicators have substantial effects on subjective status and also modify the impact of individual-level variables. However, countries with similar economic contexts might have varying educational and labour market structures, which affect opportunities and character of social hierarchies in society.

We add a new perspective to previous research by studying how country-level economic and institutional context (occupational and educational structure) impacts subjective social position in European countries. We decided to concentrate on European countries because previous research indicated that Europeans are likely to see themselves within a continental perspective (Sweeney and McFarlin, 2004). We hypothesise on the basis of the extended reference group thesis, the big-fish-little-pond and stigmatisation arguments that national educational and occupational structures modify the effect of individual-level characteristics on subjective social position. According to these arguments, the same individual-level characteristics can lead to disparate self-evaluation depending on social context. Thus, our main question is how economic inequalities, educational inequalities and occupational structures influence people’s evaluation of their social position in European countries, and how these country-level inequalities interact with the individual-level characteristics in influencing subjective social position.

This study investigates subjective social position in 21 European countries using a visual scale and data from the social inequality module of the International Social Survey Programme (ISSP) 2009.¹

Theoretical background

Subjective social position and social comparison

Studies of social stratification use subjective or objective approaches to describe people’s position in the social structure. The measurement of an objective position is often based on education level, income, occupational position or other characteristics related to occupation. Critics claim that when measured in this way, the objective social position does not specify how individuals are really ranked in society (Bottero, 2004) but how certain visible characteristics that all individuals possess are ranked (Hiller, 1973).

A subjective social position reflects people’s opinions about their location in a social hierarchy. To subjectively locate themselves, people have to perceive the extent of social inequality, identify where inequality comes from and then find their place in the hierarchy (Hout, 2008). The process by which individuals perceive their social position involves social comparison as a fundamental psychological mechanism that links objective facts to subjective evaluations (Gruder, 1977). However, subjective social position cannot be adequately understood if the role of frames of reference is ignored. Merton (Merton, 1956) developed reference group theory which is based on a concept elaborated by Hyman (1942) in social psychology. Reference group theory argues that people evaluating their own position compare themselves with other individuals or groups (Kelly, 1968; Merton and Kitt, 1950). Festinger (1954) in his theory of social comparison hypothesises that individuals prefer to compare themselves to individuals who are similar as this provides a more precise evaluation of one’s situation. People perceive a broader social world from the viewpoint of their own and their families’ and friends’ situations, and they assess their position in society in the light of the people around them (Evans et al., 1992; Evans and Kelley, 2004). Since family, friends and colleagues usually have rather similar education levels, incomes and occupations, most people see themselves as unexceptional and average. People with a high objective status perceive many people above them, and those with a low objective status perceive others as even lower. As a result, most people place themselves in the middle of a social hierarchy (Evans et al., 1992; Kelley and Evans, 1995; Savage, 2000).

The extended reference group thesis

Merton and Kitt (1950) stated that people use not only others in their own groups but also ‘non-membership reference groups’ to assess their circumstances and position. Originated from this claim, there were significant modifications and extensions of Festinger’s similarity hypothesis in subsequent decades (Fahey, 2010; Suls et al., 2002). Later research suggests that there is a great deal of diversity in the referent others with whom comparisons are made (Delhey and Kohler, 2006; Stiles and Kaplan, 2004; Suls et al., 2000). There are also systematic variations in the selection of comparison standards (upward vs downward comparisons; individual vs aggregate comparisons; comparison to the national norm; selection of specific reference groups) (Suls and Wheeler, 2000). Klein (1997) shows that aggregate information influences self-evaluation more than do individual objective characteristics. In addition, social norms shifted from the national to transnational level (Whelan and Maitre, 2009a) provide guidelines for evaluation.

However, discussions about international reference groups are rather new. The question is whether people take circumstances in other countries into account when evaluating their own position (see Delhey and Kohler, 2006; Fahey, 2007; Whelan and Maitre, 2009b, 2013). Some authors suggest that globalisation and the growth of the mass media expanded people’s horizons and understandings so that they no longer have restricted reference groups for social comparison (Rose, 2006; Schor, 1998). People are more informed than in the past and may make comparisons across countries (Fahey and Smyth, 2004). Also the economic and political integration of European societies might result in European-wide reference groups used by people in their self-ranking (Fahey, 2007; Sagi, 2011).

Previous results indicate the prevalence of upward comparisons: people choose a country for comparison that is economically better off than their own (Brown and Haeger, 1999; Delhey and Kohler, 2006; Sagi, 2011). Wilkinson (2000) argues that the more unequal a society is, the more this increases a sense of inferiority. This might imply that social comparisons are made across a wider spectrum of society as well as across countries (Delhey and Kohler, 2006; Fahey, 2007). Based on these arguments, we could formulate the extended reference group thesis which alters the restricted basis on which people make social comparisons. Today’s information society has expanded their comparison beyond country boundaries as well across the whole society.

The big-fish-little-pond phenomenon and stigmatisation argument

However, the same objective characteristics can lead to different self-evaluation depending on the frame of the reference and therefore also on social context (Marsh et al., 2008). The big-fish-little-pond phenomenon² is also based on social comparison to a generalised other. According to this phenomenon the subjective social position of different occupational, income and educational groups might depend on societal context. For example, the scarcity of professionals in one country should confer higher social position for professionals than in another country where professionals are numerous.

On the other hand, the stigmatisation argument assumes that social attainment and failure are under the control of the individual themselves and that increasing importance is attached to attainment. Stigmatisation depends on the context: the important indicator is how high or low individual attainment is compared to others and to society. Therefore, low occupational or educational attainment becomes an increasingly discrediting, stigmatising attribute for individuals if the proportion of people with low attainment decreases in society (Fiske, 1998).

Previous research on cross-national differences and hypotheses

Evans and Kelley (2004) find that despite individual-level differences in socio-demographic characteristics, economic prosperity has a substantial effect on subjective social location: when the society is more affluent, people’s subjective social position is generally higher. Following the expanded reference group argument, it is possible that Europeans take a larger reference than their immediate surroundings. Previous results also indicate the prevalence of upward comparisons (Brown and Haeger, 1999; Delhey and Kohler, 2006; Sagi, 2011). It means that if somebody lives in a country everybody recognises as poor by European standards, their subjective social position should be lower than in wealthier countries because individuals will assume some of the poorer national standing within their country as part of their own social position.

Hypothesis 1. We hypothesise that economic prosperity has a significant positive impact on subjective social position.

Andersen and Curtis (2012) find support for the classic Marx and Weber arguments that class-related differences are larger if economic inequalities between social positions are pronounced in a society. This result is also in accord with the expanded reference group argument: people may be increasingly aware of the extent of inequality, and living in an unequal society might lower their subjective social position because they make comparisons across country boundaries and the whole of society (see also Wilkinson, 2000). Previous findings demonstrate that income tends to have a much stronger effect on subjective social position in societies with a high level of income inequality compared to societies that are more equal (Andersen and Curtis, 2012). If income inequality between different social groups is high, people are more likely to perceive social differences. Also, according to stigmatisation arguments, the impact of income should vary between countries with different levels of economic inequality because in countries with the largest economic differences people belonging to the lowest level of income distribution might feel especially discredited due to larger social distance from others.

Hypothesis 2. We hypothesise that income inequality has a significant negative impact on subjective social position. We assume that country-level income inequality increases the strength of the relationship between household income and subjective social position.

Parallel to income inequality, a considerable degree of educational inequality might emphasise the educational differences between people. In other words, distances between people might be larger due to contrasting educational attainment. One mechanism that refers to the importance of educational inequality is stigmatisation: less educated people might perceive their position as especially low in the social hierarchy when educational inequality is large. In addition, the significance of education depends on how the qualifications gained are valued in the labour market and in society in general. Formulating the following hypothesis, we refer to the big-fish-little-pond argument: if one has higher education in a country with higher educational inequality (i.e. small ‘pond’), that should imply a higher subjective position than where one has higher education where educational inequality is small (i.e. big ‘pond’).

Hypothesis 3. We expect those with higher levels of education to be more likely to perceive their social position as high when educational inequality is high. At the same time, we expect less educated people to perceive their social position as lower when education inequality is high.

Previous research has not directly addressed the impact of the country-level occupational structure on subjective social position. We do not predict any main effects of occupational structure. However, we argue that based on the big-fish-little-pond argument, the proportion of higher occupational groups in the country should modify the effect of professional position on subjective social status. On the other hand, the impact of the proportion of unskilled workers might follow the stigmatisation argument. Low occupational attainment becomes an increasingly discrediting and stigmatising attribute for individuals in societies where there are a few unskilled workers (Fiske, 1998).

Hypothesis 4. Based on the big-fish-little-pond argument, we expect professionals to have lower subjective social position in societies which have a high proportion of the upper occupational group because people estimate their position more closely to the average if they know many people with a similar position to theirs. In line with the stigmatisation argument, we expect unskilled workers to have higher subjective social position in societies where the proportion of unskilled workers is high because they are not as stigmatised as they are in societies where the proportion of unskilled workers is low.

The threat of unemployment might lower the subjective social standing of individuals because it makes all people feel vulnerable and increases uncertainty in society. A rise in the national unemployment rate might indicate an economic downturn, which might have a significant negative impact on individuals’ subjective social position (Evans and Kelley, 2004). However, the national unemployment rate is not a clear indicator of economic downturns because in some countries the unemployment rate is higher during economic growth than in other countries experiencing economic downturns. A more direct measure is the growth of the unemployment rate. In societies where unemployment rises very quickly, people might be pushed down the social hierarchy because of insecurity.

Hypothesis 5. We assume growing unemployment has a negative effect on subjective social position. However, we think that people who are unemployed when the unemployment rate is rising rapidly might perceive their unemployment as temporary and therefore be more positive about their social position than those who are unemployed during more stable conditions.

Data and methods

Individual-level data

We used data from the ISSP, which is an annual programme of cross-national collaboration. Our analysis used social inequality data from the 2009 survey. We included 21 European countries in the analysis and our sample size was 24,597 individual respondents. As our main aim is to study the impact of contextual structural variables on subjective social position, we decided to restrict our analysis to European countries because previous analyses indicated that Europeans are more likely to see themselves within a larger continental perspective than would Japanese or Americans see themselves within a global perspective (see, for example, Sweeney and McFarlin, 2004). Table 1 shows descriptive information of the countries included in the analysis.

Table 1.

The mean of subjective social position for each country and values for national-level variables.

Country	Individual-level data	National-level data						Sample size
	Mean subjective social position^a (standard deviation)	GDP per capita^b	Income inequality (Gini)^c	Educational inequality (Gini)^a	Share of upper professionals and managers^b	Share of unskilled workers^b	Growth of the unemployment rate 2007–2009^b	Total N = 24,597
Finland	6.505 (0.064)	115	0.255	0.188	29	12	1.3	851
Norway	6.109 (0.039)	176	0.222	0.123	20	7	0.7	1422
Austria	6.006 (0.043)	125	0.271	0.113	17	15	0.4	1019
Sweden	5.894 (0.046)	119	0.225	0.159	25	8	2.2	1107
Belgium	5.858 (0.046)	118	0.247	0.146	33	11	0.4	1100
Denmark	5.767 (0.041)	123	0.265	0.193	22	13	2.2	1498
Switzerland	5.709 (0.045)	144	0.302	0.167	26	9	1.2	1216
Cyprus	5.497 (0.050)	100	0.293	0.188	19	18	1.4	951
Poland	5.266 (0.048)	61	0.294	0.140	22	19	−1.4	1257
Great Britain	5.238 (0.058)	111	0.357	0.120	30	11	2.3	870
Spain	4.985 (0.041)	103	0.321	0.219	21	17	9.7	1167
Estonia	4.939 (0.056)	64	0.311	0.149	28	10	9.1	979
Slovenia	4.933 (0.049)	87	0.234	0.169	24	12	1.0	990
France	4.895 (0.031)	108	0.286	0.161	22	13	1.1	2742
Czech Rep.	4.865 (0.045)	82	0.249	0.098	17	6	1.4	1199
Slovakia	4.860 (0.047)	73	0.234	0.100	16	9	0.1	1124
Croatia	4.634 (0.049)	64	0.276	0.148	17	18	0.9	1146
Latvia	4.466 (0.053)	51	0.365	0.144	26	17	11.1	1038
Portugal	4.422 (0.067)	80	0.340	0.309	16	19	1.7	950
Hungary	4.041 (0.048)	65	0.261	0.144	22	11	2.6	1008
Bulgaria	3.974 (0.053)	44	0.352	0.174	19	17	−.1	963

Source: ^aCalculations based on ISSP (2009) data; ^bEurostat database; ^cStandardised World Income Inequality Database.

Dependent variable: Subjective social position

Subjective social position was measured using a 10-box display from bottom to top. The question (Smith, 1986) was formulated as follows: In our society there are groups which tend to be towards the top and groups which tend to be towards the bottom. Below is a scale that runs from top to bottom. Where would you put yourself now on this scale? This question differs from the traditional class identification measures because it is comparable across cultures, avoids forcing respondents to choose one of several specified categories and eliminates highly politicised terms such as ‘working class’ and ‘middle class’ (Evans and Kelley, 2004; Goldman et al., 2006).

Individual-level independent variables

One aim of the analysis was to identify how contextual variables modify the effect of individual-level characteristics on subjective social position. The level of education indicates the highest level that the respondent has achieved and is recorded as follows: (1) below upper secondary education, (2) upper secondary completed, (3) above upper secondary (other qualification than university) and (4) university degree completed. Labour market status describes the respondent’s current position in the labour market. This variable distinguishes between whether the respondent is working as a manager or a professional (ISCO-88 groups 1-2), a lower white-collar worker (ISCO 3-5), a skilled worker (ISCO 7-9) or an unskilled worker (ISCO 6, 9) or if the respondent is unemployed, a student or retired. Household income is measured in quartiles, which show the respondents’ relative position in the distribution of incomes in their country. Other independent variables included in the analysis are gender and age (also the quadratic term of age).

Country-level characteristics

We analysed several national-level social and economic factors that might affect subjective social position. In order to describe the occupational structure in the country, we included the proportion of upper professionals and managers and also the proportion of unskilled workers from the total workforce. In addition, the change in the unemployment rate between the years 2007 and 2009 demonstrates recent economic growth or downturn in the country. The level of economic prosperity is measured as the percentage from the European average GDP per capita in 2009 (measurement unit of GDP per capita is Purchasing Power Standards or PPS-Euros). All these measures were obtained from Eurostat databases.

Based on the ISSP database, we calculated the Gini coefficient for educational inequality using years of schooling. This measure shows how spread people are in terms of years of schooling, but it does not indicate stratification or other qualities of the education system. We measured income inequality using the Gini coefficient (based on disposable household income data). Gini measures were obtained from the Standardised World Income Inequality Database (Solt, 2009).

Method

We used hierarchical linear models to estimate the effect of different individual- and country-level contextual variables on subjective social position. First, we calculated an empty model with no explanatory variables to predict subjective social position. The aim was to see how much variance is at the country level. After that, we calculated a model with individual-level variables. This model served as the baseline comparison for all succeeding but more complicated contextual models.

The next step was to analyse the effect of contextual variables on subjective social position. Our aim is to find out how much each contextual variable affects how people estimate their social position. Therefore, we composed a separate model for each contextual variable, and for each of these models, we also added cross-level interactions based on our hypothesis. The models are as follows, also including individual-level variables:

Model 1: economic prosperity (GDP)

Model 2: income inequality (Gini coefficient)

Model 3: Model 2 + interactions with income

Model 4: educational inequality

Model 5: Model 4 + interactions with education

Model 6: occupational structure

Model 7: Model 6 + interactions with professionals and unskilled workers

Model 8: Change of unemployment rate

Model 9: Model 8 + interaction with unemployed

The interaction terms that were statistically significant are illustrated in Figure 2, where we present the predicted probabilities for different educational, occupational and income groups over the range of contextual variables (other variables were set to the mean). Only the intercepts are assumed to be random in all models. In Appendix 1, we added two models: (1) including all contextual variables and (2) including all contextual variables as well as all cross-level interactions. This appendix shows the effects of contextual variables when controlling for country-level variables.

We measured the improvement of each contextual model in comparison with the simpler individual-level model. We used likelihood-ratio test and reported the remaining unexplained country-level variances. We also added Akaike information criterion (AIC) and Bayesian information criterion (BIC) values.

Results

Descriptive overview of subjective social position in Europe

Table 1 shows the mean subjective social position for each country and other national-level data. Descriptive statistics show that people’s estimation of their position in society varies to some extent between European countries – the mean of the subjective position ranges from 3.97 in Bulgaria to 6.50 in Finland. In general, we find Nordic countries have the highest average estimation as well as Austria and Belgium, while people living in post-socialist countries and in Portugal have the lowest opinion of their social position. There might be two different explanations to such large differences between countries. First, subjective social position measures location in social hierarchy as much as subjective well-being. Second, respondents use extended reference group comparison when evaluating their social position.

Table 1 shows significant variance in inequalities, economic development and occupational structure between European countries. Figure 1 presents the mean of the subjective social position by national-level characteristics for each country. This figure demonstrates that people have higher subjective social position in countries which have higher GDPs. Subjective social position tends to be higher in countries with less income inequality, such as Finland and Norway, and lower in countries with greater income inequalities, the most extreme examples being Bulgaria and Latvia. In addition, it seems that less educational inequality is related to higher subjective social position (e.g. Austria, Norway and Belgium), although this trend is not as clear. In particular, post-socialist countries with relatively little educational inequality and low subjective position deviate from the general picture.

Figure 1.

Relationship between national-level data and average subjective social position in the country, descriptive statistics.

Figure 1 also suggests a positive relationship between the percentage of upper professionals and perceived social position. However, Austria and Norway, which have high average subjective social position and quite low levels of upper professionals, do not fit with this tendency. In a similar way, people’s estimation of their social position seems to be higher in countries where the proportion of unskilled workers is low. Important exceptions are the Czech Republic, Slovakia, Hungary and Estonia, where the average subjective social position is low despite the small proportion of unskilled workers. No clear relationship exists between subjective social position and unemployment change. However, subjective social position is relatively low in countries where the growth in the unemployment rate was the highest between the years 2007 and 2009 (i.e. Spain, Latvia and Estonia). Although descriptive findings at the country level reveal general trends, the estimation of social position is made at the individual level and therefore, the objective position of the individual and possible interdependence of the national-level factors are taken into account in the following multi-level analysis.

Multi-level analysis of subjective social position

As the first step of the multi-level analysis, we estimated the empty model (Table 2), which showed that about 15 percent of the variance in the evaluation of social position is between countries (intra-class correlation). Including individual-level variables increased slightly the country-level variance, meaning that the subjective social position appears more different across countries after controlling for individual income, education, labour market position, age and gender. Thus, the composition of countries in terms of people’s objective social status characteristics differs. Appendix 1 shows that individual-level characteristics have a significant impact on subjective social standing: higher income, education and occupational position increase subjective social position, while the unemployed have the lowest estimation of their location in society.

Table 2.

Estimates from hierarchical linear regression models predicting subjective social position in 21 countries: effect of economic prosperity, educational and income inequality, standard errors in parentheses.

	Empty model	Individual-level model	Model 1: economic prosperity	Model 2: income inequality		Model 3: income inequality + interactions		Model 4: educational inequality		Model 5: educational inequality + interactions
Individual-level variables	−	+	+	+		+		+		+
GDP			1.58*** (0.25)
Gini for income				−7.12**	(2.82)	−7.71***	(2.86)
Income × Gini for income (ref. 1st quartile × Gini)
2nd quartile × Gini						0.14	(0.72)
3rd quartile × Gini						1.74**	(0.72)
4th quartile × Gini						1.41*	(0.73)
Gini for education								−1.50	(3.03)	0.03	(3.07)
Education × ed. Gini (ref. university degree × ed. Gini)
Below secondary × ed. Gini										−2.17***	(0.64)
Upper secondary × ed. Gini										−1.19*	(0.72)
Lower tertiary × ed. Gini										−0.71	(1.07)
Country-level variance	0.148^a	0.163	0.064	0.129		0.129		0.161		0.160
Likelihood-ratio test^b	−	−	0.00	0.01		0.00		0.62		0.01
AIC	93,489	87,993	87,973	87,989		87,985		87,995		87,988
BIC	93,513	88,155	88,143	88,160		88,188		88,165		88,183
Number of individuals	24,597	24,597	24,597	24,597		24,597		24,597		24,597
Number of countries	21	21	21	21		21		21		21

Source: ISSP (2009).

GDP: gross domestic product; AIC: Akaike information criterion; BIC: Bayesian information criterion.

Intraclass correlation.

Compared to individual-level model, prob. > χ².

p < 0.10; **p < 0.05; ***p < 0.01; ed. Gini – Gini for education.

As a next step, we included each country-level variable separately in the model with individual-level characteristics. The idea is to control separately how much each societal-level factor affects the way people estimate their social standing. Moreover, the aim of our analysis was also to find out whether the importance of the individual-level characteristics for estimating social position is shaped by country-level contextual factors. In order to consider this, we added respective cross-level interaction effects to each country-level variable model.³ The interactions that were significant are presented in Figure 2.

Figure 2.

Relationship between subjective social position and contextual variables by individual-level characteristics, fitted values from hierarchical linear regression.

Model 1 in Table 2 shows that economic prosperity has a strong positive effect on subjective social position, as expected in the hypothesis. Including economic prosperity in the model decreases the unexplained variance at the country level significantly. Thus, economic prosperity proves to be the main context-level characteristics affecting people’s opinions about their social position.

Model 2 tests the importance of the degree of income inequality (Table 2). It appears that a higher Gini coefficient strongly reduces people’s opinions about their social position. Likelihood-ratio test shows that including the Gini measure significantly improves the model compared to the individual-level model, although it reduces country-level unexplained variance much less than economic prosperity. However, further analysis shows that the degree of income inequality is important even if GDP is similar in countries (Appendix 1). This result contradicts the conclusions of Andersen and Curtis (2012) who find that the influence of economic prosperity diminished after including income inequality in the models. However, we only focused on European countries where differences in income inequality and economic prosperity are not as extreme as in other parts of the world.

Moreover, analysis shows that the interaction between income inequality in a country and individual income is significant (Model 3). In line with the hypothesis as well as previous results (see Andersen and Curtis, 2012), the subjective social position differences between income groups are smaller if general income inequality is low (also Figure 2). It appears that increasing levels of income inequality in a country particularly lower the self-estimation of people belonging to the lowest income quartile.

Model 4 shows that the extent of educational inequality has no effect on subjective social position (Table 2). We conducted an additional analysis to see whether the effect of the educational Gini coefficient is important when economic prosperity and income inequality are controlled for, but no significant influences were found (Appendix 1). Nevertheless, we cannot rule out the possibility that other measures of national-level differentiation in education could be important for estimating social position in European countries. However, the results of Model 5 show that the relationship between individual educational level and subjective social position is moderated by the educational inequality in the country (Table 2). Figure 2 reveals that especially people with low educational levels are more pessimistic about their location in society when educational inequality is high. Our explanation for this finding is simple: if educational inequality is high, people with little education feel stigmatised and, thus, they are more likely to position themselves at the bottom of the social hierarchy. On the other hand, educational inequality does not affect the subjective social position of people with higher levels of education. Thus, the findings only partly confirm our hypothesis.

Model 6 indicated that the occupational structure in the country does not have direct influence on people’s estimation of their social position (Table 3). Moreover, all model parameters show that the model with measures of occupational structure is not significantly better from the individual-level model. However, Model 7, which also includes interactions between labour market position and occupational structure, is a better fit than the individual-level model. It appears that the relationship between the labour market position of an individual and their perceived social position depends on the proportion of managers and professionals in a country (Table 2). Contrary to expectations, managers and professionals do not have lower social standing in societies where the proportion of upper professionals is high (Figure 2). Interestingly, the higher percentage of professionals, the higher the subjective social position of other people (this is why the cross-level interaction between upper professionals and the proportion of professionals is negative in Model 7). It might be that a higher proportion of upper professionals indicates better general working and living conditions for all employed people in European countries. One reason could be the composition of the group ‘other’. In societies with a low proportion of upper professionals, ‘other’ is more likely to include people who have had an ambition to become an upper professional but have not achieved this because these jobs are less available due to the occupational structure. Contrary to our expectations, unskilled workers do not have a higher subjective social standing in societies where the proportion of unskilled workers is high.

Table 3.

Estimates from hierarchical linear regression models predicting subjective social position in 21 countries: effect of occupational structure and change of unemployment rate, standard errors in parentheses.

	Empty model	Individual-level model	Model 6: occupational structure		Model 7: occupational structure + interaction		Model 8: change of unemployment rate		Model 9: change of unemployment rate + interaction
Individual-level variables	−	+	+		+		+		+
Percent of professionals			−0.04	(0.03)	−0.04	(0.03)
Percent of unskilled			0.03	(0.03)	0.04	(0.03)
Professionals × percent of professionals					−0.01**	(0.01)
Unskilled × percent of unskilled					0.01	(0.01)
Change of unemployment rate							−0.05	(0.24)	−0.05	(0.24)
Unemployed × change of unemployment rate									0.00	0.865
Country-level variance	0.148^a	0.163	0.138		0.139		0.154		0.154
Likelihood-ratio test^b	−		0.13		0.08		0.24		0.50
AIC	93,489	87,993	87,993		87,992		87,993		87,995
BIC	93,513	88,155	88,171		88,187		88,164		88,174
Number of individuals	24,597	24,597	24,597		24,597		24,597		24,597
Number of countries	21	21	21		21		21		21

Source: ISSP (2009).

AIC: Akaike information criterion; BIC: Bayesian information criterion.

Intraclass correlation.

Compared to individual-level model, prob. > χ².

p < 0.10; **p < 0.05; ***p < 0.01.

A change in the unemployment rate in the comparison years 2007 and 2009 had no effect on the subjective social position (Model 8 in Table 3). This result for European countries differs from previous studies, which found that economic downturns may decrease subjective self-placement in the social hierarchy. The reason could be that the survey was conducted in 2009, at which time it was not clear how extensive or long-lasting the economic difficulties and the related threat of unemployment in European societies would be. Thus, the effect of uncertainty on subjective social position might still have been low in those European countries experiencing the highest growth in the unemployment rate. Contrary to our expectations, Model 9 shows that the increase in the unemployment rate does not interact with employment status, so unemployed people have a rather low estimation of their social standing in spite of national-level trends in unemployment (Table 3).

Conclusion

In many European societies, concerns about rising economic inequalities are increasing. As the European Union aims towards greater social and economic cohesion, its expansion promoted further debates about social inequalities within and between its countries. In this context, it is important to understand people’s own views about their positions in the social hierarchy and relate this to how the expansion of European Union might have affected reference groups that people use for social comparisons. This study explored how economic and educational inequalities in society and occupational structure affect people’s perception of their social position and how these country-level factors modify the way individual characteristics influence subjective social position. Previous studies (e.g. Andersen and Curtis, 2012; Evans and Kelley, 2004) neglected the role of the educational and occupational structures, but we assumed that these country-level factors are important because countries with similar economic situations might have very different institutional contexts. We explained the way people make comparisons in the societal context with an extended reference group thesis, and the big-fish-little-pond and stigmatisation arguments. We focused on Europe, where, although the countries have close economic and societal relations, there are a number of differences in the structures of the societies.

Our research indicated that most people tend to see themselves as being in the middle of the social hierarchy. However, the middle varies between countries as well as for different social groups. Contrary to the simple reference group hypothesis, our research demonstrated that both the structural context and the individual’s actual social position have a substantial impact on subjective social position. This means that in addition to the reference group forces that move individuals towards the middle of the social hierarchy (Evans and Kelley, 2004), there are socio-economic forces at work that move towards greater dispersion.

In line with previous research (see Andersen and Curtis, 2012), our evidence suggests that the economic context in society affects people’s perception of their social position. Economic prosperity has a positive effect, and income inequality a negative effect, on subjective social position in European countries. There might be two explanations for the importance of economic prosperity as an influential factor. First, people in poorer societies face several constraints in their everyday life, which might lower their subjective social position. Second, people use an extended reference to other countries as well as an upward comparison. That is, Europeans living in poorer countries are more likely to make an upward comparison with people living in more affluent countries. This result does not correspond to Stiles and Kaplan (2004), who conclude that upward comparisons are made to those people who are close to oneself either in shared personal traits, or geographical location, or socially as family or friends, and downward comparisons when comparing with the national norm. However, Stiles and Kaplan suppose that technology has broadened potential reference groups.

We found that people in societies where income inequality is high are more likely to perceive differences in their social position and to link their position to their income level. This is especially true for people in the lowest sectors of income distribution who tend to have low self-estimation if income inequality is greater in the country. This result supports the stigmatisation hypotheses: the poorest people might feel more stigmatised in an unequal society due to the large social distance from others. It seems that European countries with more equitable income distribution policies avoid the very low self-estimation of people within the lowest income sections of the society.

Our findings indicate that national educational and occupational structures affect the way people estimate their social standing. We, again, found support for the stigmatisation argument: people with low levels of education are more pessimistic about their position in society when educational inequality is high. In contrast, the proportion of unskilled workers does not affect the subjective social position of the unskilled. Thus, our research indicates that the stigmatisation argument applies to people with low education and income rather than to those working in unskilled positions.

Our findings for the importance of educational inequality and occupational structure are not in accordance with the big-fish-little-pond phenomenon. Having the highest level of education in a country with large educational disparities does not increase subjective social position compared to more equal countries. In addition, people working as managers and professionals have high opinions of their positions in society whatever the general occupational structure. On the other hand, a higher proportion of professionals does seem to increase the subjective social standing of all other social groups except the professionals themselves. One reason could be that people who wish to secure higher positions are less likely to do it in societies with a small proportion of upper professionals, due to the small number of these positions in the occupational structure.

Our results suggest that rising income inequality and increasing educational inequality might increase the polarisation of subjective social hierarchies. It is important to note that higher inequalities do not improve the subjective social position of people with more resources, but lower the perceived social standing of people with the lowest income and education. The implications of this polarisation are lower levels of happiness, social trust or subjective health, which might affect cohesion of the entire society.

Unfortunately, there were no direct questions in ISSP about what kind of reference groups respondents used to compare their own social position. It means that we did not directly investigate cross-national comparisons. However, the impact of macro-level indicators on subjective social position seems to indicate that people, when evaluating their social position, are comparing themselves with others as well as aggregate data sources. We believe that these findings support the extended reference group thesis. Subjective social position is also affected by social standards across several countries.

Some limitations of this study should be recognised. Our analysis focused on the impact of social structure on subjective social position, but we recognise that differences in cultures and values might also contribute to the different estimation of social position in European countries. Due to the cross-sectional design of our data, we were not able to directly analyse how the change of societal context in time affects how people perceive their social position, and whether events in an individual’s life modify this perception. We supposed the Europeanization of the reference group. Future research should indicate if reference groups have an international, rather than European, dimension. Moreover, our analysis concentrated on mean social position. More rigorous analysis of the size of the country variances, as well as quantile regressions, would be the next step. It would also be useful to use the mixed-method design in which the quantitative approach is supplemented with qualitative research to more fully explicate the alternative social comparison processes and the psychological processes involved.

Footnotes

Appendix

Appendix 1.

Estimates from hierarchical linear regression models predicting subjective social position in 21 countries, standard errors in parentheses.

	Model 1		Model 2
Individual level
Men	0.127***	(0.020)	0.125***	(0.020)
Age	−0.023***	(0.004)	−0.023***	(0.004)
Age²	0.001***	(0.000)	0.001***	(0.000)
Education (ref. university degree)
Below upper secondary	−0.949***	(0.025)	−0.617***	(0.088)
Upper secondary	−0.561***	(0.032)	−0.376***	(0.163)
Lower tertiary qualification	−0.323***	(0.032)	−0.217	(0.106)
Labour market position (ref. skilled workers)
Managers and professionals	0.471***	(0.041)	0.725***	(0.195)
Lower white-collar	0.142***	(0.035)	0.145***	(0.135)
Unskilled workers	−0.150***	(0.047)	−0.275**	(0.137)
Unemployed	−0.446***	(0.048)	−0.447***	(0.142)
Students	0.470***	(0.054)	0.470***	(0.141)
Retired	0.092**	(0.042)	0.088**	(0.137)
Income (1st quartile)
2nd quartile	0.452***	(0.029)	0.427**	(0.202)
3rd quartile	0.753***	(0.030)	0.278	(0.205)
4th quartile	1.199***	(0.032)	0.853***	(0.208)
Country level
Gini for education	−1.963	(1.796)	−2.550	(1.792)
Gini for income	−5.190**	(2.197)	−5.663***	(2.229)
GDP	1.504***	(0.258)	1.508**	(0.256)
Change of unemployment rate	−0.001	(0.023)	−0.001	(0.023)
Percent of professionals	0.039**	(0.016)	0.040***	(0.016)
Percent of unskilled	0.060**	(0.026)	0.058**	(0.026)
Interactions
Education × Gini for education (ref. university degree × educational Gini)
Below secondary × educational Gini			−2.085***	(0.549)
Upper secondary × educational Gini			−1.147	(0.987)
Lower tertiary × educational Gini			−0.664	(0.644)
Income × Gini for income (ref. 1st quartile × Gini)
2nd quartile × Gini			0.093	(0.715)
3rd quartile × Gini			1.699**	(0.724)
4th quartile × Gini			1.245*	(0.737)
Managers and professionals × percent of professionals			−0.011*	(0.006)
Unskilled × percent of unskilled			0.010	(0.010)
Unemployed × change of unemployment rate			0.002	(0.010)
Country variance	0.042		0.041
Likelihood-ratio test^a	0.00		0.00
AIC	87,974		87,965
BIC	88,185		88,257
Number of individuals	24,597		24,597
Number of countries	21		21

Source: ISSP (2009).

GDP: gross domestic product; AIC: Akaike information criterion; BIC: Bayesian information criterion.

Compared to individual-level model, prob. > χ².

p < 0.10, **p < 0.05, ***p < 0.01

Funding

ESF PRIMUS Programme for Development of the Quality of Third Level Studies project entitled Labour Market Challenges to Higher Education: Estonia in European Context.

Notes

References

Abbott

(1987) Women’s social class identification: Does husband’s occupation make a difference. Sociology 21(1): 91–103.

Adler

Epel

Castellazzo

. (2000) Relationship of subjective and objective social status with psychological and physiological functioning: Preliminary data in healthy white woman. Health Psychology 19(6): 586–592.

Andersen

Curtis

(2012) The polarizing effect of economic inequality on class identification: Evidence from 44 countries. Research in Social Stratification and Mobility 30: 129–141.

Baxter

(1994) Is husband’s class enough? American Sociological Review 59(2): 21–46.

Beeghley

Cochran

(1988) Class identification and gender-role norms among employed married women. Journal of Marriage and Family 50(3): 719–729.

Bottero

(2004) Class identities and the identity of class. Sociology 38(5): 985–1003.

Brown

Haeger

(1999) Compared to what? Comparison choice in an internation context. European Journal of Social Psychology 29: 31–42.

Buunk

Mussweiler

(2001) New directions in social comparison research. European Journal of Social Psychology 31: 467–475.

Centers

(1949) The Psychology of Social Classes. Princeton, NJ: Princeton University Press.

10.

Centers

(1953) Social class, occupation, and imputed belief. American Journal of Sociology 58(6): 543–555.

11.

Davis

Robinson

(1988) Class identification of men and women in the 1970s and 1980s. American Sociological Review 53(1): 103–112.

12.

Davis

Robinson

(1998) Do wives matter? Class identities of wives and husbands in the United States, 1974–1994. Social Forces 76(3): 1063–1086.

13.

Delhey

Kohler

(2006) From nationally bounded to Pan-European inequalities? On the importance of Foreign countries as reference groups. European Sociological Review 22(2): 125–140.

14.

Evans

MDR

Kelley

(2004) Subjective social location: Data from 21 nations. International Journal of Public Opinion Research 16(1): 3–38.

15.

Evans

MDR

Kelley

Kolosi

(1992) Images of class: Public perceptions in Hungary and Australia. American Sociological Review 57(4): 461–482.

16.

Fahey

(2007) The case for an EU-wide measure of poverty. European Sociological Review 23: 35–47.

17.

Fahey

(2010) Poverty and the two concepts of relative deprivation. Working Paper Series 10/1, July. Dublin: UCD School of Applied Social Science.

18.

Fahey

Smyth

(2004) Do subjective indicators measure welfare? Evidence from 33 European societies. European Societies 6(1): 5–27.

19.

Festinger

(1954) A theory of social comparison processes. Human Relations 7: 117–140.

20.

Fiske

(1998) Stereotyping, prejudice, and discrimination. In: Gilbert

Fiske

Lindzey

(eds) Handbook of Social Psychology. New York: McGraw-Hill, pp. 357–411.

21.

Goldman

Cornman

Chang

M-C

(2006) Measuring subjective social status: A case study of older Taiwanese. Journal of Cross-Cultural Gerontology 21(1–2): 71–89.

22.

Goodman

Huang

Schafer-Kalkhoff

. (2007) Perceived socioeconomic status: A new type of identity that influences adolescents’ self-related health. Journal of Adolescent Health 41: 479–487.

23.

Goyder

(1975) A note on the declining relation between subjective and objective class measures. British Journal of Sociology 26(1): 102–109.

24.

Gruder

(1977) Choice of comparison persons in evaluating oneself. In: Suls

NJM

Miller

(eds) Social Comparison Processes. New York: Wiley, pp. 21–41.

25.

Hiller

(1973) The subjective dimension of social stratification: The case of the self-identification question. Journal of Sociology 9(1): 14–21.

26.

Hodge

Treiman

(1968) Class identification in the United States. American Journal of Sociology 73(5): 535–547.

27.

Hout

(2008) How class works: Objective and subjective aspects of class since the 1970. In: Lareau

Conley

(eds) Social Class: How Does It Work? New York: Russel Sage Foundation, pp. 25–64.

28.

Hyman

(1942) The psychology of status. Archives of Psychology 269

29.

Jackman

(1983) Class Awareness in the United States. Berkley, CA: University of California Press.

30.

Kawachi

Kennedy

Wilkinson

(eds) (1999) The Society and Population Health Reader: Income Inequality and Health. New York: The New Press.

31.

Kelley

Evans

MDR

(1995) Class and class conflict in six western nations. American Sociological Review 60(2): 157–178.

32.

Kelly

(1968) Two functions of reference groups. In: Hyman

Singer

(eds) Readings in Reference Group Theory and Research. New York: Free Press, pp. 199–206.

33.

Kikkawa

(2000) Changes in the determinants of class identification in Japan. International Journal of Sociology 30: 34–51.

34.

Klein

(1997) Objective standards are not enough: Affective, self-evaluative, and behavioral responses to social comparison information. Journal of Personality and Social Psychology 72: 763–774.

35.

Knudsen

(1988) Class identification in Norway: Explanatory factors and life-cycle differences. Acta Sociologica 31(1): 69–79.

36.

Marsh

(1991) The failure of high-ability high schools to deliver academic benefits: The importance of academic self-concept and educational aspirations. American Educational Research Journal 28: 445–480.

37.

Marsh

(2007) Self-Concept Theory, Measurement and Research into Practice: The Role of Self-Concept in Educational Psychology. Leicester: The British Psychological Society.

38.

Marsh

Hau

(2003) Big-fish-little-pond effect on academic self-concept: A cross-cultural (26-country) test of the negative effects of academically selective schools. American Psychologist 58: 364–376.

39.

Marsh

Seaton

Trautwein

. (2008) The big-fish-little-pond-effect stands up to critical scrutiny: Implications for theory, methodology, and future research. Educational Psychology Journal 20: 319–350.

40.

Matějů

Večernik

(1999) The renewal of the middle class and its political circumstances. In: Matějů

Večernik

(eds) Ten Years of Rebuilding Capitalism: Czech Society after 1989. Prague: Academia, pp. 207–227.

41.

Merton

(1956) Continuities in the theory of reference groups and social structure. In: Merton

(ed.) Social Theory and Social Structure. New York: Free Press, pp. 335–340.

42.

Merton

Kitt

(1950) Contributions to the theory of reference group behaviour. In: Merton

Lazarsfeld

(eds) Continuities in Social Research: Studies in the Scope and Method of ‘The American Soldier’. New York: Free Press, pp. 40–106.

43.

Oddsson

(2010) Class awareness in Iceland. International Journal of Sociology and Social Policy 30(5): 292–312.

44.

OECD (2011) Divided We Stand: Why Inequality Keeps Rising. Paris: OECD Publishing.

45.

Operario

Adler

Williams

(2004) Subjective social status: Reliability and predictive utility for global health. Psychology & Health 19: 237–246.

46.

Rose

(2006) Social comparisons and social order: Issues relating to a possible re-study of W.G. Runciman’s relative deprivation and social justice. ISER Working Paper 48, September. Colchester: Institute for Social & Economic Research, University of Essex.

47.

Sagi

(2011) Determinants of satisfaction with living standards in transition societies. International Journal of Sociology 41: 55–78.

48.

Savage

(2000) Class Analysis and Social Transformation. Oxford: Oxford University Press.

49.

Schor

(1998) The Overspent American. New York: HarperCollins.

50.

Seaton

Marsh

Craven

(2010) Big-fish-little-pond effect: Generalizability and modernisation: Two sides of the same coin. American Educational Research Journal 47(2): 390–433.

51.

Smith

(1986) Internationally comparable measurement of subjective social class. Presentation at the Planning Meeting of the International Social Survey Programme, Mannheim, April.

52.

Solt

(2009) Standardizing the world income inequality database. Social Science Quarterly 90(2): 231–242.

53.

Stiles

Kaplan

(2004) Adverse social comparison processes and negative self-feelings: A test of alternative models. Social Behavior and Personality 32(1): 31–44.

54.

Suls

Wheeler

(eds) (2000) Handbook of Social Comparison. New York: Kluwer Academic Publisher; Plenum Publishing.

55.

Suls

Martin

Wheeler

(2002) Social comparison: Why, with whom, and with what effect? Current Directions in Psychological Science 11: 159–163.

56.

Suls

Martin

Wheeler

(2000) Three kinds of opinion comparison: The triadic model. Personality and Social Psychology Review 4: 219–237.

57.

Sweeney

McFarlin

(2004) Social comparisons and income satisfaction: A cross-national examination. Journal of Occupational and Organizational Psychology 77: 149–154.

58.

Vanneman

Cannon

(1987) The American Perception of Class. Philadelphia, PA: Temple University Press.

59.

Whelan

Maitre

(2009a) Europeanization of inequality and European reference groups. Journal of European Social Policy 19(2): 117–130.

60.

Whelan

Maitre

(2009b) The Europeanization of reference groups: A reconsideration using EU-SILC. European Societies 11(2): 283–309.

61.

Whelan

Maitre

(2013) Material deprivation, economic stress, and reference groups in Europe: An analysis of EU-SILC 2009. European Sociological Review 29(6): 1162–1174.

62.

Wilkinson

(2000) Mind the Gap: Hierarchies, Health and Human Evolution. London: Weidenfeld & Nicolson.

63.

Yamaguchi

Wang

(2002) Class identification of married men and women in America. American Journal of Sociology 108(2): 440–475.