Towards a Measurement of Extreme Left-Wing Attitudes

Abstract

Previous research on left-wing extremism relied largely on measurements that are unsuitable to detect and track extreme attitudes. In the absence of a clearly defined instrument for left-wing extremism, researchers based their predictions largely on indicators like voting behaviour or the left—right self-placement scale. Building on a large-scale investigation of data from the past 20 years, the paper shows that such indicators do not serve as adequate proxies, since those labelled as ‘extreme’ do not hold extreme attitudes in most cases. Analysing a variety of left-wing extremist attitudes across the past two decades shows that there is indeed a core set of attitudes that seem to resemble what we would call a traditional left-wing extremist identity. The paper suggests an instrument that can serve as a starting point for further developing an index of left-wing extremism. Finally, the potential of the extreme left is assessed using the constructed index. The results show that the estimated potential of the new measure differs substantially from conventional measurements, not in quantity but in its composition. As a result, traditional measurements fail to predict actual left-wing extremism. In particular, anti-pluralist mindsets are heavily neglected. Implications are thus discussed for Germany, and the world at large.

ACKNOWLEDGEMENTS

I would like to thank Viola Neu of the Konrad-Adenauer-Stiftung for providing access to the data for 1997 and 2007 and Richard Stöss of the Otto-Suhr-Institut for the 2003 data. Analyses for 2008 are based on the GMF surveys of the Institut für interdisziplinäre Konflikt- und Gewaltforschung at the University of Bielefeld and were funded by a foundation consortium led by the Volkswagen Foundation. I would like to thank Andreas Hövermann and Andreas Zick for granting access to it.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author.

ABOUT THE AUTHOR

Sebastian Jungkunz is research associate and doctoral candidate at the University of Bamberg. His research interests include causes and effects of populism and political extremism.

ABOUT THE AUTHOR

Sebastian Jungkunz is research associate and doctoral candidate at the University of Bamberg. His research interests include causes and effects of populism and political extremism.

Notes

1 This also separates extremism from radicalism. Radicals are described as pursuing drastic political ideas that are still within the bounds of the free democratic basic order. Extremism in contrast goes well beyond that (see March and Mudde 2005; Stöss 2007).

2 For a brief summary of the development see the work of Olsen (2007) or Bowyer and Vail (2011, 687–690). Furthermore, this is not just a German phenomenon but applies to left-wing parties all over Europe (Gomez, Morales, and Ramiro 2016, 355), as can be seen, for instance, in the case of the Socialistische Partij (SP) in the Netherlands (van Kessel 2015, 113).

3 People also seem to make a distinction between how endpoints of scales are labelled. Jordan and Ferguson (2016), for example, found out that citizens are less likely to choose an endpoint if it is labelled as ‘extreme’ compared to ‘very’, regardless of the actual direction of political ideology (liberal vs. conservative). Treier and Hillygus (2009) further argued that it is especially those citizens that think in a multidimensional ideological belief system that are likely to rate themselves as moderate due to the potential cross-pressures they are facing when having different ideological positions on those scales.

4 Therefore scholars argue that left and right need to be measured in at least two dimensions, e.g. radical vs. conservative and authoritarian vs. democratic (Eysenck 1954, 111), egalitarian vs. anti-egalitarian (equality) and anarchic vs. illiberal (liberty) (Bobbio 1996), or economic vs. cultural (Kitschelt 2004).

5 Since the GDR item is of course only applicable in Germany, it is not suitable for studies in other countries and needs to be replaced with country-specific items.

6 As data stems from various sources, the modes of data collection differ as well. The results will thus consider potential bias for construct validity.

7 Unfortunately there is no data for anarchism or autonomism. Thus it is left to further research to disentangle those concepts.

8 Correlations are slightly stronger in the upper triangle of the matrix for East Germany and slightly weaker for West Germany. The overall picture does not change though.

9 Of course it is always difficult to set definite cutoff points. In this case, attitudes were chosen that correlated with each other at least moderately (around 0.3 and above). It could be argued that anti-globalisation should also be included but as there is a lack of data in two cases and it correlates only very weakly in two more, it is left to further research to do so.

10 The reliability for available items is still good though, as alpha stays around 0.7 across years. Only in 2003 it is a little lower with a value of 0.6. For more on the importance of reliability coefficients see footnote 13.

11 The respective question can be found in the World Value Survey questionnaire. Other, more refined items are part of the left-wing and right-wing authoritarian batteries of Altemeyer (1996) and van Hiel, Duriez, and Kossowska (2006). Inter alia, they ask whether people should follow a leader who will fight the establishment or whether a revolutionary movement is justified in demanding obedience and conformity of its members.

12 The absolute level of correlation between items is not extremely high. Nevertheless, striving for this would miss a critical point. The index is designed to grasp the concept of left-wing extremism. All items are hence supposed to measure different facets, which means that inter-item correlations are not informative of the reliability of the items (see Gerber et al. 2010, footnote 2; more broadly Kline 2000, 11–15).

13 Unfortunately due to the nature of the data it is not possible to include all items in all years. The index contains variables for left-wing radicalism (socialism, nationalisation, GDR nostalgia) and anti-pluralism (dictatorship). In 2003 the GDR item was replaced with anti-capitalism to match the same amount of items in the index, since the GDR variable was not available. Comparing the shortened index to a longer six-item version returns a correlation of 0.98 in 1994, 0.97 in 2003 (both also included anti-Americanism and anti-imperialism) and 0.95 in 1997 (including anti-capitalism and elite criticism). Excluding the GDR item does not lead to a substantial change of results in any case.

14 Results for overall models from confirmatory factor analyses can be found in the appendix (see Table A3). Most models indicate good fit according to recommendations by Hu and Bentler (1999).

15 This pertains to the items per se, as well as to the aggregated index. Results for individual items are not shown here.