Comparative survey analysis using WVS data: social challenges and global value shifts |
|
| Chair | Ms Kseniya Kizilova (World Values Survey Association ) |
| Coordinator 1 | Dr Tatiana Karabchuk (United Arab Emirates University) |
| Coordinator 2 | Professor Eduard Ponarin (High School of Economics) |
Attitudinal research consistently demonstrates that electoral losers show weaker political support than electoral winners, and that the difference is larger if respondents lose repeatedly. Marginalized ethnic groups are, by definition, consistently “losing” the electoral contest, which is expected to undermine their political trust and support for the political system. But while the political science literature has often warned of such dire consequences, we lack reliable evidence of the link between ethnic group membership and attitudinal outcomes like political trust and support. This paper makes two contributions towards filling this gap: it directly examines the link with large-scale survey data, and in doing so proposes a novel operationalization of ethnic group membership that takes political relevance into account.
We test the “ethnic loser” hypothesis in an analysis of ethnic power relations and individual attitudes in 49 countries over the four latest waves of the World Values Survey (WVS), conducted between 1995 and 2014. We examine three dependent variables to capture the nuances of political support: respondents’ confidence in political institutions; their satisfaction with the way democracy works in their country; and their perceptions of electoral integrity, to be able to adjudicate whether levels of support are due to respondents’ assessment of democratic performance or indeed a lack of legitimacy of the democratic process itself.
The main independent variable is respondents’ ethnic group membership, for which we use a novel operationalization that takes political relevance into account. Previous analyses of ethnic background – often merely included as control variable, where included at the individual level at all – have used a single ethnic group identifier available in their respective survey and has done so in isolation of the political context. But ignoring the politicization of ethnicity leads to invalid inferences about the role of ethnic “loss” in increasing feelings of political inclusion or exclusion among minorities. We hence operationalize ethnic group status by combining survey data with information from the Ethnic Power Relations (EPR) dataset, which codes ethnic groups in each country with regard to their political relevance and status, and use a variety of country-specific classifications regarding ethnicity, language, religion, and region, to match the group descriptors in the EPR dataset and assign them the respective EPR status.
We examine the data in a multilevel analysis that considers both individual and contextual characteristics as well as changes over time in a single model, which allows differentiating between over-time and cross-sectional effects. Preliminary results show that marginalized ethnic groups are indeed less trusting and satisfied with the way democracy works in their country. However, with regard to perceptions of electoral integrity, ethnic differences are insignificant or insubstantial; differences in perceptions of electoral integrity are instead explained by political partisanship alone. The results suggest that substantive, policy representation is more important than descriptive representation, even in ethnically divided societies: the ethnic winner–loser gap in political attitudes is not due to a psychological “sore loser” effect after elections but to the extent to which respondents feel represented in between elections.
This paper presents a new angle on the effects of globalization on generalized trust, exploring how the relationship is mediated by patterns of inequality and immigration. Recent research has established a direct, but at times ambiguous effect of inequality and diversity on generalized trust, while there is a comparatively little research regarding the role that globalization plays in determining levels of trust. We combine these areas to examine how trust and globalization are related as well as who profits from globalization and who loses out.
Trust has a well-established relationship with inequality and with diversity. Areas with high levels of inequality tend to exhibit low levels of trust. Although there are many external characteristics to consider, the explanation is simple. Where inequality is high, competition for jobs is also high and trust levels are diminished. Likewise, when diversity is high, trust between people is hampered. The effect this has on trust should be doubly so when high levels of inequality and immigration are found together.
Conversely, the relationship between trust and globalization is less clear. On the negative side, globalization weakens social and political institutions, exacerbates inequality, and threatens the economic security of individuals. On the positive side, globalization increases economic development, trade, and international cooperation and eases the flow of goods, people, and information. This suggests that in some facets, trust is negatively affected by globalization while in other facets, it is positively affected. In other words, there are “winners and losers” in how trust is affected and we need to disaggregate how we understand globalization to understand who those winners and losers are.
Our analysis incorporates multi-level models to analyze 104 countries from six waves of the World Values Survey. In general, we find that increasing economic, political, and social aspects of globalization strengthen trust. However, when we examine the relationship separately for different social classes, we observe stark differences between social classes in the relationship between trust and globalization. Globalization increases trust in a consistent way only at the top of the class structure, particularly for economic and political aspects of globalization. The effect of all forms of globalization on trust is predominantly negative for individuals in working and lower classes. Additionally, we find that rising immigration does not affect levels of trust at the upper end of the class structure, but has a negative effect for individuals in the middle and bottom portions.
Globalization has caused unavoidable and uneven structural changes in all regions across the globe and thus it is imperative that the consequences of these changes are understood. Our findings suggest that the positive effects of the relationship between trust and globalization are only felt by those whose economic well-being is not threated by inequality, diversity, or globalization and not equally shared across the socio-economic spectrum. The consequences of this could possibly lead to greater conflict – particularly in diverse and unequal societies – and even undermine the positive features of globalization, in trust, and in all social, political, and economic outcomes.
Multinational surveys have been begun to conduct in last decades. This study compares two such surveys. Most attention is paid to the World Value Survey (WVS) that one of the oldest multinational survey but we compare it to the European Social Survey (ESS) that is about 20 years younger. Our focus is in happiness in countries where some type of transition has been happened. During last 25 years two such transition shocks can be recognized, mainly in Eastern Europe. One is the collapse of the Soviet Union and similar countries, and the other is concerned joining the European Union.
The WVS gives opportunity to analyze the influence of the first transition since one survey was made around 1990, and four others after that. The ESS cannot tell anything about that collapse since it was started in 2002 but it can be used for examining the second transition.
The topic can be looked from several perspectives but we select one, happiness after that shock. We first analyze it at aggregate level, looking for the average happiness time series’. We found that happiness varies by generations. Hence we looked at changes by birth cohorts. It is not automatic to compare these two surveys since the happiness questions, for example are not equal. In particular. We could not choose the most common happiness question “Taking all things together, how happy would you say you are?” since its scale in the WVS is. 1=’Very Happy’, 2=’Quite Happy’, 3=’Not Very Happy’ and 4=’Not At All Happy’ whereas the ESS includes 11 categories (0=’Extremely Unhappy’, …, 10=’Extremely Happy.’). These two alternatives were found to be inappropriate in comparisons. It was good that the WVS includes another questions with 10 categories, ‘Life Satisfaction’ (1 to 10) whereas the ESS with 11 categories (0 to 10). In order to improve the comparison we made a new Happiness variable from both questions. This was made by linear transformation giving a fair variable in comparisons.
The data are not available well from many countries, being best for Russia and Poland. Hence we analyzed the first shock for both of these, and the second shock for Poland. Partial information from Belarus, Slovenia, Estonia and Romania is used as well but these do not give as clear results. Russia results of the WVS show clearly that happiness declined after the collapse (to mid 1990’s) on average except for youngest cohort but then happiness increased quite evenly in all cohorts to 2006, becoming lower after that. The results for Poland are quite similar including the young age cohort but the changes in other age cohorts are more complex.
The results due to joining the EU are general fairly clear in Poland, Happiness increased but became lower quite soon. The average of the youngest cohort was very high first but did not increase essentially later.
It is not of course automatically clear whether those results are only due to that ‘transition’ shock but it is one reason.
Networks, norms and trust are among the most common terms used to define social capital (Cook 2005). In theory, social capital can have “spillover” effects (Aarstad et al.2010), where social capital in one sphere can increase participation in another. However, only limited attention has been paid to the effect of workplace social capital on membership in voluntary associations (Wilson & Musick 1997). The current paper aims at testing the “spillover” effects of workplace on formal membership in a broad cross-national perspective. More precisely, we are interested in the impact of employment status, supervision, level of autonomy, creativity at work, and cognitive or manual character of tasks. We are seeking to answer the question whether these characteristics have a universal impact on membership across nations or it is dependent on the social context.
To answer the question the current study follows multilevel approach and incorporates individual-level data from the 5th and 6th waves of the World Values Study and country-level data on associational freedom provided by Freedom House. We treat associational freedom as a proxy for civic infrastructure in the country hypothesizing that it should strengthen the influence of occupational characteristics. Hierarchical regression modeling with cross-level interaction term is used as the main method for testing our suggestions. We operationalize formal membership through participation in four types of organizations: 1) sport and recreational organizations, 2) art, music or educational organizations, 3) humanitarian or charitable organizations, and 4) environmental organizations. Individual level control variables include age, gender, educational attainment, position in a 10-point income scale, and respondent’s self-assessment of his/her social class position, generalized trust and emancipative values (see Welzel 2013).
Our analysis revealed cross-cultural variation in the impact of occupational characteristics. Supervision status has a relatively stable positive effect across different social contexts but it is stronger in the countries with low level of associational freedom. In general, this finding is in line with previous research which had suggested that high status work positively supported volunteerism in the US (Wilson and Musick 1997). At the same time, the role of other characteristics is dependent on social context and cannot be treated as universal. The positive influence of employment is statistically significant only under high level of associational freedom. The impact of self-employment is totally contextual. It decreases formal membership under low associational freedom and increases it under high level. This result could be explained by the cross-cultural differences in the employment status. Probably in less developed countries that usually have low level of associational freedom self-employed status produces risk and vulnerability while in developed countries it results in the increase of leisure time which could be spend on civic activity and achieving common goods. Autonomy at work do not related to formal civic engagement, while cognitive and creative job have robust positive impact in countries where people are free to join and create associations.