The Question of External Validity of Trust in the ESS R1-7 Data
The Question of External Validity of Trust in the ESS R1-7 Data

Summary/Abstract: The European Social Survey data collected in seven rounds (2002-2014) offers a good opportunity to gain an answer to the question about the external validity of obtained data. From a wide variety of topics included in the core module of ESS questionnaire we concentrate on social and political trust; not only for social and/or political reasons of importance of trust, but also for the explanatory power assigned to this concept across many disciplines in the area of social and behavioral sciences. Method: The sample consists of 246, 982 adult respondents in representative samples from 25 EU countries participating in the ESS, collected in the 7 biennial rounds of the European Social Survey (2002–2014). In regression analyses where social trust, trust to national institutions and trust to international institutions served as criterions, we used 3 groups of indicators as predictors: Group A. – GDPppp per capita_2014, GINI index, European Parliament turnout 2014; Group B. – Human Development Index_2014, Corruption Perceptions Index_2014, Quality of Life Index_2014; Group C. – ESS questionnaire variables - Political Participation Scale, Social Satisfaction Scale, Well–being Scale, Subjective general health (item). Results: A stable and strong correlation between economic indicators and social trust is well-documented. Our analysis confirmed that 3 indicators of Group A together are significant predictors of trust for people with GINI as significant single predictor. The complex indicators addition into the group of predictors increased powerfully the explanatory power of MODEL 2 and in this constellation the CPI manifested as the most effective predictor. The addition of four ESS indicators in this case did not change the situation dramatically and in this case the CPI manifested itself as the most powerful predictor. The general idea that there is a connection between “simple” indicators of national economic performance and trust towards political institutions in ESS data (Round 1-3) was tested already. Zmerli (2013) confirmed GDPppp as a significant predictor of trust towards institutions and it was also the result in our analysis when GDPppp, GINI and EPturnout (MODEL 1, tab.2) were used as predictors of trust towards national institutions. In the case of international institutions, the GINI index was significant (MODEL 1, tab.3), which is in accordance with the notion that economic inequality can also lower trustworthiness in international political institutions (Zmerli, van den Meer, 2017). Political trust is expected to have an impact on the propensity to vote (Grönlund & Setälä, 2007 confirmed this hypothesis in the ESS Round 1 dataset) so it seemed reasonable to expect the reverse influence, but in our analyses this expectation was not confirmed through the statistically significant correlations in both the indicators of political trust. The Complex indicators addition into the group of predictors increased the explanatory power of MODEL 2 in both cases. Surprisingly, not the Human Development Index (Listhaug, Ringdal, 2007; Zmerli, 2013) but the Corruption Perception Index together with the GINI manifested themselves as the most powerful predictors of trust towards national institutions. In the case of trust in international institutions, MODEL 2 was not significant, although the number was close to the critical level (p<0,083) and none of the predictors was statistically significant – though the CPI was close (p<0,064). Also, all predictors’ correlations in both the indicators of political trust were statistically significant. The addition of four ESS indicators (political participation, social satisfaction, well-being and subjective general health) further increased the explanatory power of the empirical model in both cases (MODEL 3, tab.2 and 3). Despite the statistical significance of predictors’ correlations in both the indicators of political trust between single predictors in MODEL 3, only social satisfaction was significant.

• Details
• Contents