Suchergebnisse

The article contributes to the issue how to deal with measurement non-variance. I address a well-established scale by Middleton (1963) which was created to measure alienation. However, unlike commonly treated in literature, there is evidence that the scale is two-dimensional, and consists of the measures of anomie and alienation. I use the data from two datasets, where the scale was most recently applied, World Values Survey (2011), and Euromodule (1999-2002), for a set of diverse countries representing Western, post-communist, and Eastern societies. The datasets are analyzed separately. Results of confirmatory factor analysis followed by multigroup comparisons give evidence that for a set of countries the two-dimensional scale is applicable or preferable. Full measurement invariance is reached for Russia and Kazakhstan in the World Values Survey, and Slovenia, Switzerland, Germany, Austria, Turkey, and South Korea in the Euromodule. Approximate measurement invariance using Bayesian statistics enabled to reach approximate scalar invariance in Russia and Kazakhstan in the World Values Survey and in Slovenia, and Switzerland in the Euromodule dataset. Additionally, to be sure that the two dimensions are indeed distinct, I used a set of indicators to predict each of the factors.

Recent studies have shown that the coronavirus pandemic not only temporarily increased support for incumbent politicians and trust in experts but also triggered an authoritarian response. Because the pandemic has significantly affected individuals' goals, needs, and control over their lives, we expect that it has generated emotional reactions. In this article, we study how concerns about COVID-19 relate to institutional trust (trust in political institutions and experts) and a preference for populist right parties—directly and indirectly—via emotions. Our theoretical framework relies on the "rally around the flag" hypothesis, the cultural backlash theory, as well as appraisal theories of emotions. We analyze a novel data set collected as part of the Longitudinal Internet Studies for the Social Sciences Panel Survey for the Netherlands at the beginning of the second wave of the pandemic (October 2020). Our findings reveal that concerns about COVID-19 are positively related to institutional trust but unrelated to preferences for right-wing populism. The relationship between concerns about the coronavirus crisis and trust in political institutions is mediated via fear and sadness; these emotions also explain opposition to right-wing populist parties. We interpret our findings in relation to research on the rally around the flag effect, right-wing populism, and emotions. We discuss the implications of our results in the context of the coronavirus pandemic and other "rally"-inducing events.

Multiple group confirmatory factor analysis has become the most common technique for assessing measurement invariance. However, higher-order factor modeling is less frequently discussed in this context. In particular, the literature provides only very general guidelines for testing measurement invariance of second-order factor models, which is a prerequisite for conducting meaningful comparative research using higher-order factors. The current paper attempts to fill this gap. First, we explicate the constraints required for identification of the invariance levels in a multiple group second-order factor model. Second, in addition to the conventional interpretation of the results of this assessment, we suggest an alternative view on the invariance properties of a second-order factor as evidence of structural rather than measurement invariance. Third, we present an empirical application of the test which builds on Seeman's alienation scale and utilizes data from eight countries collected in 2008-2009. We found empirical support for metric invariance of both the firstand second-order factors, but no support for scalar invariance of the first- and second-order factors. However, we find pairs of countries where scalar invariance for both the first- and second-order factors is supported by the data. We finalize with a discussion of the results and their interpretation.