Introduction

This is an interactive appendix, accompanying the study:

'When the Whole is Greater Than the Sum of its Parts: On the Conceptualization and Measurement of Populist Attitudes' by Alexander Wuttke, Christian Schimpf, Harald Schoen

Purpose of this App

Populism is a multi-dimensional concept that is characterized by the simultaneous presence of its concept components. Not all procedures to aggregate the concept's subdimensions into populism scores account for the defining characteristic of populist attitudes as an attitudinal syndrome. Hence, this Shiny Web Application enables users to assess the sensitivity of substantive findings to the different operationalization strategies discussed in the study.

Content

Interacting with the study's data study, you can compute various estimates of interest employing the Sartori, Goertz and Bollen operationalization strategies.

How many populists?

  • Set a threshold to compute the share of populists in various countries using different populism scales.
  • Assess the sensitivity of estimated shares of populists to different, plausible thresholds.

Populist Attitudes and Correlations

  • Plot correlations of various substantive variables of interest with populist attitudes, based on the Goertz or Bollen operationalization strategy.
  • Assess the differences between populism scales and across countries.

Internal Structure of Scales

  • Show a visual matrix of correlations between the subdimensions of populism and its composite score.
  • Assess the disparity between Goertz, Sartori, and Bollen operationalization strategies for various populism scales in different countries.

Sources and Information

  • Data Sources
  • Weights
  • Replication Material

Acknowledgement

To create this app, we drew on the insights of various contributors on Stackoverflow. These sources are listed in the replication code for this app. We also want to thank Bruno Castanho Silva and colleagues for allowing us to draw on their replication dataset.

To assign individuals into membership categories (e.g., populist vs. non-populist) the Sartori approach requires the specification of a threshold. To account for the noncompensatory relationship of concept concept, citizens (or other observational units) are only categorized as populists if they exceed the specified threshold on all subdimensions. Importantly, the level of the threshold must be justified by substantive reasoning that reflects the researcher's theoretical propositions about the concept's essential core.

Which threshold do you set to determine the share of populists in a population?

First, choose a populism scale and the country of interest.
(Thresholds may vary across dimensions. Here, we assume equal thresholds on all concept dimensions. Note that this application is not intended for comparisons of mean levels across countries.)
There are (at least) two different strategies to set tresholds, each reflecting different assumptions about the concept itself.

Set an absolute threshold using unstandardized indicators or a relative threshold using standardized indicators.
Current Choice: Absolute threshold: Populists are those who accept all tenets (subdimensions) of populism. The researcher needs to specify the intensity of the agreement necessary to qualify as a populist.
Current Choice: Relative threshold: Populists are those who agree with the tenets of populism more strongly than their fellow citizens. The researcher needs to define how much more strongly individuals have to agree compared to their fellow citizens to qualify as populists.
Which threshold do you set?
Select populism scales and countries of interest to compare the correlations between substantive variables and a) the Goertz Aggregation, b) the Bollen Aggregation, and c) the scale's discrete subdimensions.
The correlation matrix allows you to assess the strength of the correlations between the subdimensions of populism. In addition, you can assess the disparities of the composite scores derived from the Goertz, Sartori and Bollen operationalization strategy.

Overview

This page contains information on the datasets that were used to create the plots, the weights that were applied in the plots in the second panel, and information on the replication material.

Datasets

  • ANES: The American National Election Studies - 2016 Time Series Study (www.electionstudies.org). These materials are based on work supported by the National Science Foundation under grant numbers SES 1444721, 2014-2017, the University of Michigan, and Stanford University.
  • AUTNES: Aichholzer, J., Kritzinger, S., Wagner, M., Berk, N., Boomgaarden, H., Mueller, W. C. (2018). AUTNES Comparative Study of Electoral Systems Post-Election Survey 2017 (CSES Edition). doi:10.11587/W193UZ.
  • BES: Fieldhouse, E., Green, J., Evans, G., Schmitt, H., van der Eijk, C., Mellon, J., Prosser, C. (2018). British Election Study, 2017: Face-to-Face Post-Election Survey [data collection] (Version 1.3). doi:10.5255/UKDA-SN-8418-1.
  • GLES (Cross-Section): Rossteutscher, S., Schmitt-Beck, R., Schoen, H., Wessels, B., Wolf, C., Wagner, A., Melcher, R., Giebler, H. (2019). Post-election Cross Section (GLES 2017). GESIS Data Archive, Cologne. ZA6801 Data file Version 4.0.1. doi:10.4232/1.13235.
  • GLES (Panel): Rossteutscher, S., Schmitt-Beck, R., Schoen, H., Wessels, B., Wolf, C., Preissinger, M., Kratz, A., Wuttke, A., Gaertner, L. (2018). Short-term Campaign Panel 2017 (GLES). GESIS Data Archive, Cologne. ZA6804 Data file Version 6.0.0. doi:10.4232/1.13150.
  • LISS: Jacobs, K. (Researcher), van der Brug, W., van der Kolk, H., van Klingeren, M., Lehr, A., van der Meer, T. (Contributors) (2017). LISS Panel. Election Survey Ukraine referendum. 151.3 Measurement 3. Single Wave Study administered by CentERdata (Tilburg University, The Netherlands). See www.lissdata.nl.
  • Replication Dataset: Castanho Silva, B., Andreadis, I., Anduiza, E. Blanusa, N., Corti, Y. M., Delfino, G., Rico, R., Ruth, S., Spruyt, B., Steenbergen, M.R. Littvay, L. (2018). 'Public Opinion Surveys: A New Scale.' In: The Ideational Approach to Populism: Concept, Theory, and Analysis, eds. K. A. Hawkins, R. Carlin, L. Littvay and Cristobal R. Kaltwasser. London: Routledge. See https://populism.byu.edu/Pages/Appendix.

Weights

Where applicable (and available), we applied weights for the plots shown in the section 'How many populists?' These weights are:

  • ANES: Design weight to adjust for sampling design (Variable: V160102)
  • BES: Post stratification which was applied on top of selection weights (Variable: wt_demog_cses)
  • GLES: Design weight to correct for oversampling in East Germany (Variable: w_ow)

Replication Material

The replication material for the paper and this app are available from: https://doi.org/10.7910/DVN/KPS1KY