Tuesday 16th July
Wednesday 17th July
Thursday 18th July
Friday 19th July
Download the conference book
Survey effects in secondary analysis of pooled data |
|
| Convenor | Mr Cristiano Vezzoni (University of Trento) |
| Coordinator 1 | Mr Ferruccio Biolcati-rinaldi (University of Milan) |
The growing availability of secondary data offers the possibility to tackle the same research question with different studies (Eurobarometer, ESS, EVS, EU-SILC, ISSP, etc.) that share similar operationalization and sample designs. Pooling different studies allows to increase the number of cases and expand the period of analysis, while controlling for data quality. Nonetheless, the pooling procedure introduces the potential for biases due to survey effects. Survey effects may depend - as known - on several factors (Glenn 2005, 43-50): data collection mode (face-to-face, telephone, self-administered); sampling designs; training and supervision of interviewers, coding procedures (these constitute the so-called house effect), question wording, question ordering and, more generally, the topics covered by the questionnaire (questionnaire context effect). In literature on substantive topics, different strategies to deal with survey effect are proposed, but a systematic reflection is lacking notwithstanding the relevance of the problem.
The session welcomes papers that address substantive research questions using pooled surveys data from different studies, with a focus on strategies adopted to detect and deal with survey effects.
Purpose of the study. This study deals with the creation of an instrument for measuring environmental concern. Data. The data of the WVS study for 15 European countries and ISSP survey for 9 European countries allow to build two one-dimensional cumulative scale of mobilization for the defense of the environment.
Methods. The Mokken Scale Analysis is applied to confirm the reliability and one-dimensionality of the scales, which respond to the strongest level of the scales according to Mokken's definition. The analysis of Guttman Errors permits to isolate particular patterns of activism.
Results. The validity for construct in relation to previous literature has been analyzed by observing the trend of the indexex in the sub-groups. Such procedure has also allowed to identify additional aspects on which to investigate. In particular, Holland and Great Britain show particular patterns that might constitute specific ways to relate to the environment. The environmental concern indexes developed using 2000 ISSP and WVS datasets enable the isolation of an often elusive phenomenon: the presence of individuals with a pronounced degree of activism but who do not share the attitudes of the majority of environmentalists. Further, the results underline how, with the sole exception being the level of education, the structural variables are scarcely correlated with environmentalism, while cultural factors play a decisively more important role.
Conclusions. The indexes developed represent useful instruments for measuring environmental concern. Finally, some remarks regarding the sampling procedure and the wording of the questions are added.
Studying policy congruence between citizens and their representatives in the European Parliament (EP) necessitates information on the position of various political actors: voters, national political parties, and candidates to the EP or supranational political parties. Although high quality comparative studies have been designed to overcome problems caused by cross-nationality, there are still some important discrepancies that deserve attention. Existing research on the EP relies on positions of political actors measured using various methods of data collection (e.g. voter/candidate/expert survey, manifestos and voting advice applications), all with their merits and limitations.
The question that arises is to what extent these measures are comparable and compatible. In this paper, we argue that choices among different measures depend on the specific research question at hand. We discuss the theoretical and methodological trade-offs involved, as well as the differences and similarities of various measurement techniques used in the literature. Then, we visually demonstrate how political actors' policy positions compare to each other on the basis of different measures. Finally, we use the concordance correlation coefficient to measure compatibility using common denominators. We conclude with remarks regarding measure suitability for specific types of research puzzles and the levels of compatibility among them.
The question of the rise and fall of political trust in Western democracies, its relevance, causes and effects, continues to remain high on the research agenda in comparative political science. Thanks to the emergence of regional surveys over the past decades, researchers now have abundant data to test the applicability of theories related to political trust in a comparative perspective. However, such analyses rest on the assumption of the cross-national comparability of the survey items. I seek to facilitate research on political trust in a comparative perspective that includes the non-Western realm by scrutinizing the measurement equivalence of the concept of political trust in Asian, African and Latin American democracies. In order to do so, I first merge survey data from the Afrobarometer, Latinobarometro and the Asian Barometer. In a second step, I analyze the cross-national configural, metric and scalar invariance of political trust by applying multigroup confirmatory factor analysis. In a third step, I reflect on the results in terms of possible survey effects. The analysis therefore contributes to insights regarding the possibilities and limitations of using cross-national survey data on political trust.