Introduction
The past three decades have seen a significant expansion in the scope and authority of international organisations (IOs), such as the United Nations (UN), the World Bank, and the World Trade Organization (WTO).Footnote 1 This trend has paralleled an increase in the number of societal groups that publicly debate and contest these organisations, raising fundamental questions about the processes through which IOs become visible and controversial in domestic public debates.Footnote 2 In recent years, such questions have gained an increasingly prominent position on the research agenda of a burgeoning literature on the politicisation of IOs. IO politicisation, commonly defined as the process through which an IO becomes salient and controversial at the level of mass politics,Footnote 3 is brought about by two sets of factors. A first body of research focuses on institutional factors, emphasising how the growing political authority of IOs, for example through majority decision-making and judicial power, constrains domestic governments and makes IOs more visible among citizens.Footnote 4 A second strand of literature centres on attitudinal factors, suggesting that the increasing spread of transnational competenceFootnote 5 and cosmopolitan attitudesFootnote 6 enhances citizens’ ability to understand, critically assess, and contest IOs’ policymaking.
While attitudes are central to this latter strand of research, we know little about the cognitions underlying these attitudes.Footnote 7 The most basic of these cognitions is citizens’ awareness of international policymaking. Understanding why citizens learn more about IOs than others, and how they convert the information they acquire into knowledge, is central to the theory and practice of the politicisation of IOs.Footnote 8 To begin with, the politically aware are typically male, white, well-educated, and financially secure,Footnote 9 making knowledge a ‘resource whose distribution favours some groups and interests while disadvantaging others’.Footnote 10 In the context of global governance, the imbalance in terms of citizen awareness affects the degree to which different parts of the citizenry can grasp and participate in communicative processes leading to the politicisation of IOs.Footnote 11 Furthermore, at the societal level, the differences in awareness of international politics between societal groups have implications for the flow of information about IOs’ policymaking in societies and the susceptibility of mass publics to elite manipulation.Footnote 12 Finally, the unequal distribution of awareness among mass publics may have implications for the visibility of citizen interests and, in turn, for the equality in representation of these interestsFootnote 13 through societal groups acting as a ‘transmission belt’ between the global citizenry and IOs,Footnote 14 such as non-governmental organisations (NGOs),Footnote 15 corporations,Footnote 16 social movements,Footnote 17 and self-appointed spokespersons.Footnote 18
This article asks what factors account for the variability in citizen awareness of IOs. Awareness is defined as the extent to which an individual’s knowledge is extensive and organised.Footnote 19 Since awareness has significant conceptual and empirical overlap with related concepts, such as political sophisticationFootnote 20 and knowledge,Footnote 21 I use them interchangeably.Footnote 22 Drawing from previous research on political psychology, political interest, and political equality, this article argues that people’s economic background and cosmopolitan identity shape knowledge of IOs. While there is evidence for a range of factors that affect knowledge about domestic politics, this article uses the concept of attitude importance to explain why economic factors and cosmopolitan identity are central predictors of knowledge in the specific context of IOs. To summarise the argument in short, individuals who attach personal importance to an attitude are especially knowledgeable about the attitude object. This article puts forward that an attitude toward an IO becomes important for mainly two reasons: first, if people perceive that IOs’ policymaking to affect their economic well-being; and second, if people identify with groups or individuals worldwide, whose lives may be affected by the IO.
Three hypotheses are derived from this argument and examined through a statistical analysis of a specific indicator of citizen awareness of IOs – the degree of factual knowledge about the UN Security Council. As discussed in more depth in the research design section, factual knowledge has advanced as our best measure of political awarenessFootnote 23 or knowledge.Footnote 24 The focus on the UN Security Council is motivated by three reasons. First, the Council is the only IO for which survey data on a factual question are readily available. Second, the Council is potentially the most powerful IO,Footnote 25 being ‘endowed with tremendous formal power by the UN Charter’.Footnote 26 Its decisions can be expected to be relatively visible in domestic public debates, making it reasonable to assume that there is considerable variation in awareness across different societal groups instead of a skew towards a small group of citizens holding knowledge about this central institution in world politics. Third, despite a large body of research on the Council,Footnote 27 we know little about citizens’ awareness of the politics of the Council, which is why this article fills an important lacuna in the literature.
The statistical analysis is based on a unique dataset including individual-level and country-level variables based on survey data for 17 Asian and European countries.Footnote 28 The results of a series of multilevel regression models of citizen awareness of the Council strongly endorse the developed hypotheses and suggest that: wealthier citizens are more aware of the Council; income inequality depresses awareness among relatively poor people in most of the countries that are being studied; and identification with groups or individuals across borders provides an impetus for citizens to acquire knowledge about the Council. I conclude that many citizens may be excluded from communicative processes that politicise the Council owing to economic factors and social identity, with implications for the ‘knowledge gap’ between different socioeconomic status groups.Footnote 29
This is not to say that these are the only factors shaping individuals’ knowledge of IOs – just that how much citizens know about IOs has mostly to do with their economic background and cosmopolitan identity. If attitude importance is in fact a cognitive mechanism improving knowledge about IOs, it leads us to expect that economic factors and cosmopolitan identity are crucial explanatory factors for citizen awareness of IOs, since these factors affect how important people perceive the IO to be for their lives. However, apart from these factors, a number of control variables included in the regression models have explanatory power. These alternative explanatory factors are also typical predictors of knowledge of domestic politics: gender, education, and political institutions. The concluding section of this article discusses the implications of these findings taken together for the study of the politicisation of IOs.
The antecedents of citizen awareness of IOs
Political knowledge is commonly conceived as cognitively stored, correct information about political facts.Footnote 30 Existing theories of political knowledge rest on the premise that people cannot retain all available information in memory owing to the complexity of the social world and personal limitations in time and cognitive resources.Footnote 31 Hence, people must be selective in their learning. To understand why some citizens learn more about IOs than others, we must consider the causal process through which people become motivated to seek information about various topics. A central mechanism raising people’s motivation to learn about specific political issues or objects is attitude importance.Footnote 32 Attitude importance refers to a person’s subjective interest in and concern about an attitude. People usually know when they care about an attitude towards a political issue or object, and they know when they are not especially interested in one. Attaching personal importance to an attitude leads people to use the attitude to process information, thereby acquiring higher levels of knowledge.Footnote 33
Consequently, we can expect people who think of their attitudes toward an IO as relatively important to be more interested in the IO and to gather more information about it. Building on insights from the political psychology literature, I argue that the reason why people might attach personal importance to their attitudes towards IOs is twofold. First, citizens may perceive the attitude toward an IO to be instrumental to their own privileges, material or physical well-being. All domestic audiences have potentially something to gain from the policymaking of IOs, yet some pay more for it than others.Footnote 34 For example, individuals who hold capital that can be invested internationally may benefit more from increases in trade and capital flows pushed for by IOs with authority in the area of trade, and may therefore attach importance to their attitudes toward IOs such as the WTO. Second, an attitude toward an IO may become personally important due to social identification with groups or individuals worldwide, whose lives may be directly affected by the policymaking of an IO.Footnote 35 For example, individuals may attach personal interest to their attitudes toward IOs engaging in peacekeeping missions, sanctions or military interventions, such as the UN Security Council, if they perceive these IOs to affect either their own well-being or that of groups they identify with. The remainder of this section develops hypotheses about how economic conditions and social identification shape knowledge about IOs through the causal mechanism of attitude importance.
Economic context
Theories of political knowledge typically assume in the tradition of Anthony Downs (1957) that knowledge about the political world is costly, imposing considerable cognitive demands.Footnote 36 Starting from this assumption, students of political knowledge and interest have significantly advanced our understanding of how people’s economic resources shape the level and accuracy of general political knowledge.Footnote 37 In the domain of global governance, individual wealth may spur people’s interest in IOs for two reasons.
First, given that political knowledge requires cognitive and economic resources, poorer people may have fewer resources to devote to attending to and elaborating information about IOs than the more affluent. Assuming they make decisions about gathering knowledge just as they make decisions about other goods,Footnote 38 they should exhibit lower levels of knowledge about IOs. This assumption appears warranted in international politics, as learning about IOs is likely to be even more costly than learning about domestic political institutions. Citizens rely predominantly on national news media to learn about international politics given the absence or weakness of a global public sphere.Footnote 39 However, national politics is more prominent in national news coverage and people rarely have access to rich and varied information and debate about IOs. Hence, they may incur greater costs to accumulate similar levels of knowledge about international political institutions than about domestic political institutions.Footnote 40
Second, wealthier people may be able to influence the political agenda to a greater extent than poorer people, precluding international political issues from being publicly debated that threaten richer people’s privileges while potentially benefitting poorer people. Wealthier people may have incentives to remove discussions about international policy from the political agenda that may challenge their interests, limiting the scope of the public debate to those political issues that are relatively innocuous to richer people, shaping the attentiveness and demands of poorer people.Footnote 41 As Robert A. Dahl puts it, ‘a political issue can hardly be said to exist unless and until it commands the attention of a significant segment of the political stratum’.Footnote 42 Thus, the ‘nondecision-making power’Footnote 43 of the more affluent may shape poorer people’s interest in international politics in such a way that they accept their role in the existing order of things and lose interest in their attitudes towards IOs.Footnote 44 The less wealthy may be confronted with a political system that constantly fails to develop alternatives regarding international political issues that may be beneficial to them, leading them to believe that their circumstances are unchangeable, natural, or without alternative, thereby losing interest in their attitudes towards IOsFootnote 45 and in becoming engaged in international politics.Footnote 46
To illustrate, economic reforms supported by the International Monetary Fund (IMF) benefit segments of domestic societies in different ways. IMF-sponsored policy reforms have been found to increase income inequality and to decrease labour’s share of income, even after controlling for a country’s wealth.Footnote 47 Whether such reforms increase welfare by benefiting relatively poor people or not depends not only on IMF conditionality, but also on the particular characteristics of domestic policy reforms. Domestic political considerations will largely determine who will bear the burden of restructuring the economy,Footnote 48 and the choice of policy instruments is often influenced by the political power of various income groups,Footnote 49 indicating that wealthy citizens have incentives to oppress conflict-prone aspects of IMF involvement in the public debate.
To provide an example from the area of international development, there is evidence that technical assistance programmes financed by IOs, such as the Asian Development Bank, the European Union, the UN, and the World Bank, that aim to promote market liberalisation, have in some cases not prevented governments or ruling elites to move backward from the achieved level of a market-based economy, favouring a relatively small and wealthy elite.Footnote 50 Again, IOs’ financial involvement may yield distributional benefits for the relatively wealthy and politically powerful, who, in turn, have incentives to prevent public contestation of these benefits.
In sum, we would expect richer people to have a keener interest in learning about IOs than poorer people, translating into the following hypothesis.
Hypothesis 1. The higher a person’s income in the distribution of national income, the more that person will be aware of IOs.
The expectation that individual income and knowledge are systematically related directs us towards a potential conditioning effect of income inequality. Taking cues from a number of studies on political interest and equality, we would expect the concentration of economic resources to affect the link between income and knowledge about IOs.
There are two complementary reasons for the ramifications of economic inequality on the political interest among poorer citizens.Footnote 51 First, as interest and engagement in politics involves costs,Footnote 52 fewer resources are available to less affluent people to seek information about IOs the more economic resources are concentrated in a society. Second, poorer citizens may feel consistently powerless in societies where only a small, economically or politically powerful elite controls access to and debates about political information, which stymies the political interest of poorer people.Footnote 53 A greater concentration of wealth and power should therefore imply that poorer people repeatedly find that international political issues debated are not of interest to them, and conclude that it is not worth the effort to be concerned about IOs. Taken together, we would expect income inequality to depress knowledge about IOs especially among poorer people. This logic gives rise to Hypothesis 2:
Hypothesis 2. The relationship in Hypothesis 1 is magnified when income inequality is relatively high.
Cosmopolitan identity
Furthermore, citizens may be more interested in an IO if that organisation’s policymaking impinges on the lives of individuals or groups they identify with. Footnote 54 For example, residents of a high-income country that does not receive multilateral development aid may nonetheless see the relevance of IOs such as the IMF or the World Bank for reference groups or individuals in poorer countries they identify with, some of which may benefit from multilateral aid. Likewise, citizens of countries that have experienced a long period of peace may become personally invested in their attitudes towards an IO operating in the area of international security, such as the UN Security Council, as they identify with residents of countries that are witnessing military conflicts and that are affected by UN policy, such as peacekeeping operations.
Social identification with groups or individuals in other countries or world regions often coincides with a more general cosmopolitan orientation.Footnote 55 While the two concepts are overlapping, the latter is more encompassing. Cosmopolitanism is a phenomenon shared by people that have humanity as a whole as a relevant identity group, adhere to universal ethics, and have attitudes towards people from other places and cultures that are rooted in globalism, universalism, and humanism.Footnote 56 Cosmopolitanism thus encompasses both social identification and broader as well as more stable core values.Footnote 57 Socialisation theory suggests that a cosmopolitan identity results from an individual’s set of relatively stable symbolic predispositions formed early in life, through interactions with family, friends, neighbours, or teachers.Footnote 58 However, some people may gradually modify or completely change their social identification after their formative years, for example as a consequence of repeated inter-personal experiences with transborder social relations arising from immigration, travel or job mobility.Footnote 59 By implication, an effect of cosmopolitan identity on knowledge about IOs is based on individuals’ relatively stable predispositions, in contrast to the supposed effect of economic conditions, which is owing to the power imbalance between wealthier and poorer people.
In sum, attitude importance is conceived as the mechanism bringing about an effect of social identification with groups or individuals across borders that may be affected by an IO on knowledge about that organisation. This logic is summarised in the final hypothesis:
Hypothesis 3. People with a cosmopolitan identity are more likely to be more of IOs.
Research design
The argument developed above suggests that the determinants of individuals’ knowledge of IOs exist on two levels: an individual level and a national or macro level. At the individual level, income and social identification affect the level of awareness of IOs. Higher levels of income (Hypothesis 1) and a cosmopolitan identity (Hyothesis 3) should be positively correlated with higher knowledge levels, and the positive effect of income should be magnified at greater levels of income inequality (Hypothesis 2). Income inequality thus provides a context factor within which individuals exist and which shapes the impact of income on awareness of IOs.
To examine the explanatory power of these hypotheses, one would, ideally, like to have data for a series of factual questions about different IOs to capture as much variation in citizen knowledge about IOs as possible. However, data on factual knowledge about IOs is only available for the UN Security Council. This article therefore uses individual-level data from the ASES and a range of cross-national datasets offering measures for country-level factors that may affect people’s knowledge about the Council. As the ensuing sections discuss in detail, these cross-national data are derived from the Quality of Government DatasetFootnote 60 and merged to the ASES dataset. The resulting dataset is cross-sectional, covering 17205 individuals in 17 Asian and European countries in 2001, and allows for analysing the relative power of individual- and country-level factors in explaining knowledge about the Council.Footnote 61
Operationalisation
The appropriate measurement of political knowledge has been the topic of a great deal of scholarly research. Building on Philip E. Converse’s seminal work,Footnote 62 previous studies have shown that knowledge of political facts is the most useful indicators for people’s political knowledge about specific facts and their political awareness more generally.Footnote 63 Importantly, John Zaller conducts a range of experimental tests and a thorough discussion of how well different measures, including factual political knowledge, perform as indicators of political awareness.Footnote 64 Zaller deems a person’s summary score across a series of factual tests of public affairs knowledge as the appropriate measure for political awareness, but emphasises that measures of this kind capture general political awareness.Footnote 65 Unfortunately for this study, a multi-item measure of factual political knowledge about the UN Security Council is unavailable. The ASES provides a single-item question tapping knowledge about the Council. Awareness is measured using the responses to the question of whether respondents know the names of any of the five countries with permanent seats on the UN Security Council.Footnote 66 I distinguish between three groups of citizens: those who have the lowest level of awareness about the Council, those holding middle levels of awareness, and those with the highest level of awareness. The allocation of citizens across three categories of awareness allows me to examine what factors determine to which category citizens belonging to, rather than drawing conclusions about what factors contribute to a decrease or increase in awareness. The variable awareness takes on a value of 0 if the respondent did not know any of the countries with a permanent seat, gave incorrect answers, or knew one country with a permanent seat in the Council. Answers are coded 1 if two or three countries with a permanent seat are correctly named, and 2 if four or five countries are correct.Footnote 67
The fact that this analysis relies on a single-item measure of knowledge about Council politics begs the question of whether individuals who are knowledgeable about the specific aspect of Council politics captured by the measure are habitually attentive to other aspects of Council policymaking as well. Although this may not be the case for all citizens, the variation in this measure provides an indication of more profound differences in citizens’ awareness about the Council. In this respect, Converse emphasises that general, chronic political knowledge ‘measures must be carefully constructed and multi-item, but it does not take much imagination to realize that differences in knowledge of several … “minor” facts are diagnostic of more profound differences in the amount and accuracy of contextual information voters bring to their judgments’.Footnote 68 In the context of this study, we would expect the indicator awareness to primarily capture knowledge about the permanent member states of the Council, and to give some indication about how much people know about UN Security Council policymaking more generally.
Figure 1 depicts the variation in the dependent variable. Importantly, Figure 1 reveals that awareness of the Council is particularly low in the Philippines, Singapore, and Thailand, but also in European states such as Greece, Ireland, and Portugal. Interestingly, Figure 1 does not provide evidence that citizens would be more aware of the Council, on average, in countries with a permanent seat in the Council, France, and the UK. Furthermore, it is interesting to note that some countries with borders to countries that were the targets of UNSC resolutions at or before the time the survey was conducted exhibit a relatively large share of citizens in the high awareness category, while others do not. Among the 17 countries included in the data set, three had borders with countries whose politics were directly affected by UN Security Council by way of a resolution: Greece, whose neighbour country Cyprus was the target of several Council resolutions, and Indonesia and Malaysia, as East Timor was the subject of UNSC resolutions in the end of the 1990s.Footnote 69 Yet while both Indonesia and Malaysia have a similar share of people in the high awareness category, Malaysia has a much larger share of citizens in the low awareness category than Indonesia. By contrast, in Greece, fewer people are the high awareness category and more in the low awareness category than in Indonesia and Malaysia. To explain the variation in awareness among citizens that live in different countries, the regression analysis presented in the ensuing section will include several country-level variables whose operationalisation is discussed below.
Citizen awareness about the UN Security Council. Note: N=17205. Darkest shade depicts the percentage share of the lowest level of awareness, second darkest shade depicts middle levels of awareness, and lightest shade depicts highest levels of awareness.
Hypothesis 1 requires the measurement of relative income. I code the variable income by dividing the country-specific samples into income quintiles to capture a person’s income relative to other people’s income in the same country.Footnote 70
To operationalise Hypothesis 2, I include a variable income inequality that represents an estimate of the Gini index in equivalised (square root scale) household disposable income.Footnote 71 Higher scores on this variable indicate greater income inequality. The most unequal countries in the sample are Thailand, Malaysia, Philippines, and Singapore, whereas the most equal countries are Sweden, France, Germany, and South Korea. When looking at the scores of the other countries in between, Portugal, Indonesia, and the UK are fairly unequal economically, whereas Japan, Ireland, and Taiwan are fairly equal.
To operationalise Hypothesis 3, I use a question about whether respondents identify with a larger group that includes people from other countries, for example, as European, Asian, or others, such as global communities. The resulting variable cosmopolitan identity is dichotomous and takes on a value of 0 if citizens do not think of themselves in this way and 1 if they do. The idea behind this variable is that people identifying with groups of people or individuals from other countries or cultures – and perhaps even with an all-inclusive group of all human beings – are more likely to identify with a group that is affected by the policymaking of the UN Security Council.
The regression models control for a number of factors that previous literature has shown to shape knowledge of domestic politics. To begin with, age is a continuous variable ranging from 18 to 79. In view of the ongoing discussion about gender differences in political knowledge,Footnote 72 gender enters the regression analysis as a dichotomous variable equal to 1 for females and 0 for males. Moreover, a person’s sense of and concern about the degree to which an IO may affect her or his own life may also be influenced by the formal education of that person.Footnote 73 Longer education beyond the pre-adult stage may lead to greater knowledge about multilateral institutions owing to increased exposure to readings and informal discussions about international matters.Footnote 74 Once individuals have acquired prior political knowledge, it will be easier for them to understand new political information and to store it in long-term memory.Footnote 75 Consequently, and following previous analyses of the relationship between educational attainment and political knowledge,Footnote 76 education is operationalised through two indicators. First, I include a dichotomous variable that is 1 if a person completed secondary education and 0 if otherwise. Second, I use a dummy variable that equals 1 if the respondent has completed postsecondary education and 0 if otherwise.
The country-level controls are derived from Jan Teorell et al. (2011). To identify these factors, I take cues from the literature on political interest and engagement. To begin with, I include a measure of the size of electoral districts. District magnitude and political interest are positively related because larger districts generate a greater proportionality of the votes and therefore a greater sense of citizen control over politics.Footnote 77 Consequently, I consider the possibility of an effect of district magnitude on awareness by introducing a variable district magnitude that captures the standard magnitude of the average district in the lower house.Footnote 78 Higher numbers indicate larger districts. Furthermore, I include a measure of the degree of pluralism of the party system. Previous research argues that more political parties lead to greater uncertainty about responsibility for politics and, in turn, to lower levels of political engagement.Footnote 79 In a similar fashion, greater uncertainty about which political parties are responsible for political decisions may discourage people to seek more information about IOs as well. Hence, I use a variable party pluralism that represents the count of the largest party’s number of seats divided by a legislative assembly’s total number of seats, expressed in fractions.Footnote 80 Higher numbers indicate less pluralist party systems. Last, citizens may have more incentives to engage politically if they live in presidential and not in parliamentary systems, since the division of power between the legislative and the executive in presidential systems may increase citizens’ access points in the policymaking process. Likewise, federalism may increase citizens’ opportunities for political access and therefore encourage political engagement.Footnote 81 In this study, the type of political system in a country is represented using a variable parliamentarism representing parliamentary systems (coded 2), semi-presidential systems (coded 1), and presidential systems (coded 0). In the data set, the Philippines is the only country with a purely presidential system. France, Indonesia, South Korea, Singapore, and Taiwan are coded as semi-presidential systems. A second indicator variable for the type of political system in a country represents the degree of unitarism on a scale from 0 to 2. The most federal countries in this study, scoring 0 or 0.5, are Germany, Malaysia, and the Philippines, whereas the most unitary countries in the data set, scoring 2 on the index, are Greece, Indonesia, Ireland, South Korea, Portugal, Singapore, Sweden, and the UK.Footnote 82
Regression method
The relative effects of the independent variables on the dependent variable, awareness, are estimated through a series of multilevel models that adjust for the correlation of the error components at the individual and country level. As the dependent variable awareness takes on three ordered values, the multilevel models will estimate a latent variable as a linear function of the independent variables and two cut points. The main model is written as follows:Footnote 83
$$\eqalignno{ & logit\left\{ {Pr(y_{{ij}} \,\gt\,s\!\mid\!{\bf x}_{{ij}} ,{\rm }\rzeta _{{1j}} )} \right\}={\equals}\beta _{2} (Income_{{ij}} ){\plus}\beta _{3} (Income\,\,{\rm }inequality_{j} ) \cr & {\plus}\beta _{4} (Income_{{ij}} {\times}Income\,\,{\rm }inequality_{j} ){\plus}\beta _{5} (Cosmopolitan\,\,{\rm }identity_{{ij}} ){\plus}\beta _{w} {\mib W}_{{ij}} {\plus}\beta _{z} {\mib Z}_{j} {\plus}\rzeta _{{1j}} {\plus}{\varepsilon}_{{ij}} , $$
where Pr(y ij >s|x ij , ζ1j ) is the cumulative probability that respondent i living in country j has a level of awareness UN Security Council politics that is higher than the threshold s; W are vectors for the individual-level controls; Z are vectors of country-level controls; ζ1j is the random intercept of the cumulative logits that varies over countries j, which is included because survey responses are likely to vary across national contexts; and ε ij is the error term that is separate for country-level j.
Empirical results
Before discussing the results from the multilevel models, I analyse the extent to which differences in awareness are owing to the fact that citizens live in different country contexts. An analysis of variance (ANOVA) model that does not include any independent variables reveals that the random-intercept variance is estimated as 0.19, implying that about 5 per cent of the variation in the level of awareness lies at the country level (see Table 1).Footnote 84 Unsurprisingly, this percentage is low given that the data are individual-level data, but indicates nonetheless that country-level factors should be included in the regression models to explain this part of the variation in knowledge about Council politics.
ANOVA model predicting awareness.
Note: Estimation from random-intercept multilevel ordered logit regression model, with states as second-level unit.
Robust standard errors in parentheses. Significance level: **p<0.01. Number of individuals: 17205. Number of countries: 17.
Table 2 reports the results from the multilevel analyses. In models 1–3, I test the relative explanatory power of the three hypotheses separately. Models 4–6 as well as 7–9 replicate models 1–3 by adding the individual-level and country-level controls, respectively. All hypotheses are tested concomitantly in model 10.Footnote 85
Multilevel models predicting awareness of the UN Security Council.
Note: Estimated odds ratios from random-intercept multilevel ordered logit regression models, with states as second-level unit. Robust standard errors in parentheses. Significance levels: *p<0.05, **p<0.01. Number of individuals: 17205. Number of countries: 17.
I begin by discussing the findings with regard to Hypothesis 1. The results from the models testing for a direct effect of income on awareness (models 1, 4, and 7) strongly endorse Hypothesis 1. An increase in an individual’s income in the national income distribution appears indeed to matter for how much that individual knows about the Council. Comparing two citizens that score similarly on all variables but one of which belongs to the next higher income quintile, the citizen belonging to the higher income quintile has a 17 per cent greater odds of holding middle levels of awareness as compared to low levels of awareness (or high levels of awareness as compared to low or middle levels of awareness).
The evidence for Hypothesis 2 is mixed. Hypothesis 2 is tested through a product term between income and income inequality. Figure 2 depicts this product term graphically on the basis of the estimations in model 10, showing how the statistical significance and magnitude of the marginal effect of income on awareness changes at different levels of income inequality.Footnote 86 The solid line shows the estimated marginal effect and the dashed lines indicate the bounds of the 95 per cent confidence interval for this estimate. The marginal effect of income is significant in almost all countries except for the most equal country in the sample, Sweden, suggesting that income inequality depresses knowledge among the poor in countries with intermediate levels of income inequality. This implies that at medium levels of inequality, inequality decreases knowledge among the poor, while stimulating knowledge of the rich, tying in with previous literature demonstrating that inequality depresses political interest and participation especially among the poor.Footnote 87 However, Figure 2 also shows that if inequality is very high (in Malaysia, the Philippines, Singapore, and Thailand), the magnitude of the – still positive – effect of income decreases. This suggests that in highly unequal societies, richer people still know more than poorer people, but the effect of income becomes weaker. This latter finding does not conform with the expectation expressed in Hypothesis 2, but ties in with previous literature arguing that inequality at very high levels might depress interest among the poor, and that rich people in very unequal societies do not have equal interest in IOs than rich people in medium unequal societies. As conflicts with poorer people about international political issues decrease, the more affluent have fewer incentives to engage politically as well.Footnote 88
Effect of income on awareness of the UN Security at levels of income inequality. Note: N=17205. Graphical depiction of interaction term based on estimations from model 10.
There is supportive evidence for Hypothesis 3. As the models including the indicator for cosmopolitan identity show, awareness is positively associated with cosmopolitan identity. A person with a cosmopolitan identity has 46 per cent greater odds of holding more knowledge (see model 3). This result is in line with previous studies emphasising that a feeling of belonging to an all-inclusive worldwide group of people is an important predictor of public opinion towards international politics.Footnote 89
With regard to the alternative explanations, the results suggest that men know more than women, reflecting a usual finding in the political knowledge literature.Footnote 90 However, the survey instrument to measure awareness of IOs may, as it is not based on gender-relevant items, gloss over important variation in awareness of the UN among men and women.Footnote 91 Holding a degree from secondary or postsecondary education appears to improve knowledge levels, whereby the effect of postsecondary education is particularly large. A citizen having accomplished secondary education has a more than twice the odds of holding higher levels of awareness than a citizen without a degree from secondary education. Having a degree from postsecondary education has more than four times the odds of being more aware of the UN Security Council (for example, model 4). The larger coefficient of postsecondary education suggests that people with postsecondary education may live in social and professional environments that facilitate learning about international politics. This finding ties in with previous accounts of knowledge about domestic politics.Footnote 92 Finally, there is evidence that political institutions may affect people’s awareness of the Council. Previous literature has found people in systems with greater district magnitudeFootnote 93 and federal institutionsFootnote 94 to have a keener interest in politics. There is some evidence that people know more about the Council if they have experiences with relatively high district magnitude and federal institutions. Moreover, parliamentary institutions and awareness of the Council are positively associated. This contradicts previous evidence about a negative relation between parliamentary institutions and political interest and participation.Footnote 95 That the estimate for parliamentarism presented here does not conform to expectation may be due to the fact that among the 17 countries in the sample, only one is presidential and six are semi-presidential. This measure may not capture important variation in awareness of the UN in other presidential systems owing to its skewed distribution. Hence, re-estimations of the models on other data may yield more convincing conclusions with regard to an effect of parliamentarism on political knowledge about IOs.
Robustness of the results
The findings presented in Table 2 are robust across a range of model specifications. First, all models that include the interaction term were replicated by including income as a random coefficient to allow for the possibility that the effect of income on awareness is not the same in different countries. The results remain robust (see Appendix, Table B3).
Second, all models were replicated by including a measure for media exposure. News media may be a source of information about the politics of the Council,Footnote 96 but is positively correlated with and may hence be affected by income (r=0.177) and postsecondary education (r=0.215), respectively, and is therefore not included in the main regression table. The indicator measures the frequency of unintentional news media exposure and is based on two questions. The first question yields a dichotomous measure that equals 1 if respondents use the Internet and 0 if otherwise, whereas the second question yields a dummy variable that is coded 1 if the respondent receives an international satellite or cable television. The resulting variable is an additive index of these two questions that ranges from 0 to 2, assuming that people who use the Internet and receive international satellite or cable television are more likely to be incidentally exposed to international news. The results indicate that news media exposure increases awareness of the Council, while the remaining results remain robust to this change in model specification (see Appendix, Table B4).
Third, the variable party pluralism may have biased the estimates presented in Table 2 owing to its moderately high correlations with district magnitude (r=–0.302) and parliamentarism (r=0.185), respectively. However, replicating all models by excluding party pluralism does not change the interpretation of the main regression models (see Appendix, Table B5).
Finally, two additional country-level measures are included to explore whether a country’s participation in or exposure to the policymaking of the Council shapes citizens’ knowledge about the Council. These measures are weakly or moderately correlated with nearly all country-level variables (see Appendix, Table B2 and are therefore not included in Table 2. The first measure captures the degree of country participation in the Council. It is coded 2 if a country is a permanent member of the Security Council (France and the UK), 1 if it was a temporary member in the year in which the survey data was collected (Ireland and Singapore), and 0 otherwise.Footnote 97 The coefficient of participation in the Council is significant in most models and in the concomitant model 10, and all other results remain robust (see Appendix, Table B6). Substantively, these findings indicate that the more a country participates in the decision-making process of the Council, the more citizens of this country might know about the Council: citizens from countries that are relatively more involved in the Council have a 23 per cent higher odds to know more about the Council (see model 10).
The second country-level measure captures citizens’ exposure to the Council’s decision-making. Residents of countries that have been the target of Council resolutions or have sent troops during a UN peacekeeping operation may know more about the Council. The UN Security Council resolutions decided upon during the ten years prior to the survey provided information about countries that were affected by military conflicts or that had borders with countries where military conflicts prevailed.Footnote 98 As discussed above, Cyprus was the issue of several resolutions, which might have affected Greece. Furthermore, the engagement of the Council in East Timor might have affected Indonesia and Malaysia. Consequently, the measure exposure to the Council’s authority is coded 1 for Greece, Indonesia, and Malaysia, and 0 for the rest of the countries in the sample. The results corroborate the descriptive findings in Figure 1, suggesting that citizens in these three countries do not display a clear pattern where relatively many citizens have middle or high levels of awareness. However, the coefficient for exposure to the Council is statistically significant in some models, including the concomitant model 10, which might be diagnostic of an effect of exposure to the Council’s authority on knowledge levels among some citizens.
Conclusion
To sum up, this article examines the sources of citizen awareness of IOs on the basis of evidence from the UN Security Council. Using an original dataset on 17 Asian and European countries, the article demonstrates that wealthier citizens are more aware of Council politics, and that income inequality depresses awareness among poorer people in the greater part of the countries that are being studied. Moreover, there is evidence that identification with a global community provides an impetus for individuals to acquire knowledge about the Council. With respect to the alternative explanations, the results tie in with previous literature focusing on domestic politics, suggesting that males and the better educated,Footnote 99 as well as people who live in systems with relatively large voting districts Footnote 100 and federal institutions, know more about the Council.Footnote 101 Furthermore, the unequal distribution of political power among the Council’s member states and states’ exposure to the Council’s policymaking may shape citizen awareness of the Council.
How do these results contribute to our understanding of citizen awareness of other IOs? The UN Security Council is a central body in global governance with a mandate in international peace and security, whose policymaking bears heavily on development and aid. The scope of Council politics and the robustness of the results presented in this article suggest that the findings should be taken seriously in debates on public awareness and the sources of politicisation in global governance. However, it would be premature to endorse conclusions about the sources of citizen awareness of IOs other than the Council based on the evidence at hand. Rather, this analysis is meant as one step forward toward a research agenda that examines citizen awareness of IOs comparatively or in the context of other IOs. Two main limitations to the generalisability of the findings to other IOs apply, each at the same time indicating fruitful avenues for future research.
First, the degree to which citizens become knowledgeable and potentially involved in the public contestation of IOs may vary across issue areas and IOs. To begin with, people may identify with groups or individuals abroad that are affected by IOs in different ways. Consequently, they may attach more or less importance to their attitudes toward an IO. Furthermore, income may not raise citizen awareness of all IOs in the same way. For example, the privileges and wealth of richer people may depend more on the policymaking of IOs with authority in the area of trade than on IOs with authority in the area of security, with implications for richer people’s incentives to influence the public debate about IOs, and poorer people’s knowledge levels.
Second, countries’ representation in and exposure to specific IOs may matter for citizen awareness. The unequal allocation of power among the Council’s member states and the differences in these countries’ exposure to the Council’s authority are found to have implications for citizen knowledge about the Council. Likewise, we would expect that IO-specific variables in the context of other IOs have explanatory power. A natural next step for future research would be to examine whether the theoretical argument put forward in this article holds in the contexts of other IOs and other issue areas as well, controlling for indicators of countries’ particular relationships with IOs.
Taken together, this article suggests two specific avenues for future research on the role of citizens and citizen awareness in the public contestation of IOs, which would shed light on who is knowledgeable, participates, and is represented in public debates on IOs. A first useful step would be to tie the study of cosmopolitanism more closely to the study of awareness and skills in the global citizenry. Cosmopolitanism is commonly referred to as a combination of cosmopolitan identity and values.Footnote 102 Many people understand political issues in terms of values. When elaborating knowledge, forming attitudes, and making choices, they base their opinions on the connections that they draw between political issues and core beliefs.Footnote 103 Future studies could explore how connections between values and international political issues affect citizen awareness of international politics.Footnote 104 Second, the findings in this article underline the importance of studying when and why poverty and income inequality affect citizen representation or participation in debates about IOs. Poorer citizens are unlikely to be a part of the trend towards greater public awareness about and scrutiny of IOs, especially in countries where wealth is relatively unequally distributed. This trend may be exacerbated if economic inequality was to rise among households in the future, as it has in both the developing world and in advanced industrialised countries over the past three decades .Footnote 105
Appendix A: Survey questions
Awareness: Five countries have permanent seats on the Security Council of the United Nations. Can you tell me the names of any of these five countries?
Cosmopolitan identification: Some people also think of themselves as being part of a larger group that includes people from other countries, for example, as European, Asian, Chinese, Islamic, etc. How about you, do you think of yourself in this way? (‘European’, ‘Asian’, ‘Chinese’, ‘Islamic’, ‘Other – Specify’, ‘No, I do not think of myself in this way’).
Secondary education and postsecondary education: What is the highest educational level you have attained?
News media exposure: For each of the following, could you please tell me whether or not it applies to you. (‘Applies’, ‘Does not apply’) – I use the Internet at home or school/work – I receive an international satellite or cable TV service.
Appendix B
Summary statistics.
Correlations between independent variables.
Robustness checks including income as a random coefficient.
Robustness checks including news media exposure.
Robustness checks excluding party pluralism.
Robustness checks including participation in the Council.
Robustness checks including exposure to the Council.
Acknowledgments
Earlier versions of this article were presented at the ECPR Joint Sessions of Workshops in 2014 and at the Global and Regional Governance Seminar at Stockholm University. I am grateful to the participants at these events and to the anonymous reviewers for helpful comments. I thank Jon Kuyper for the many insightful discussions we had. The research for this article was funded by the European Research Council (grant DII 200971) and the Nereus programme, which is a collaborative research programme by the Nippon Foundation, the Stockholm Resilience Centre, Duke University, the University of British Columbia, and three additional partners.
