Authoritarianism on the left and the right: A sequel

In an earlier blog post, I discussed the problems with a new scale developed by Conway and colleagues to measure left-wing authoritarianism. I wrote that a better scale for measuring this construct could benefit research in psychology. Shortly after this, another team of researchers (Costello et al., 2021) published a new left wing authoritarianism scale. Although I have not studied this scale myself, it looks more promising and lacks the obvious problems of the Conway et al. scale. It is most likely an improvement on previous measures of left-wing authoritarianism.

Nevertheless, the paper written by Costello et al. on their new scale is disappointing in some respects.

A misleading narrative

It perpetuates a sensationalist (and polemical) narrative that was later picked up by the Atlantic and Quilette. According to this narrative, there is a widespread and remarkably biased belief among psychological researchers that authoritarianism on the left does not exist, and therefore it has not been studied. In the introduction of their paper, Costello et al. write: “Many decades later, a core question remains unresolved: are some individuals on the left disposed to authoritarianism?”

Although it is possible to find a handful of extravagant quotations on the rarity of left-wing authoritarianism in the history of the field, authoritarianism scholars have rarely denied the existence of left-wing authoritarianism. In fact, a plethora of research programs have studied authoritarianism on the left for decades, particularly in formerly Communist countries in Eastern Europe.

A persistent empirical finding is that authoritarianism is, other things equal, more common among right-wing conservatives than others, even when it is measured in an ideologically neutral way. This is the conclusion I drew (together with John Jost) in a recent review (Nilsson and Jost, 2020) of research on the association between authoritarianism and political ideology. Costello et al. falsely attribute the claim that left-wing authoritarianism is “‘the Loch Ness Monster’ of political psychology” to us. Nowhere, in this paper did we deny that authoritarianism on the left exists or that it is common some contexts. We even discussed left-wing authoritarianism in Eastern Europe and South East Asia. As I wrote in my previous blogpost, left-wing authoritarianism “is common in contexts where the socially sanctioned authorities and norm systems are left leaning. ” There is no conflict between any of these assertions and the observation, made by Popper (1945) among others, that totalitarian ideologies on both the left and the right pose a threat to the open society.

Conceptual confusion

In fact, the sort of research presented by Conway et al. and Costello et al. (and many others), which operates within a standard variable-oriented paradigm, has no relevance at all to the question of whether “some individuals on the left are authoritarian”. Statistical analyses of this sort address associations between variables–they cannot make existence proofs. The notion that constructing a scale to measure something proves that it “exists” (in some unspecified sense) is based on naive ontological ideas. Psychological constructs are abstract mathematical idealizations. Personality scales measure idealized quantities of a statistical abstraction–the “average person”.

This research is also not relevant to the question of whether a “shared psychological core” underlies authoritarianism on the left and on the right. To validly address this question, it is necessary to construct an ideologically neutral measure of this core and demonstrate that it is measurement invariant across left- and right-wingers who score highly on it. Thereafter, the extent to which ideology moderates the associations between the core of authoritarianism and other phenomena can be studied to map similarities and differences between manifestations of authoritarianism on the left and the right.

Left- and right-wing authoritarianism scales may of course be useful for many other purposes. But if we want to investigate the association between authoritarianism in general and ideology, then we need neutral scales that are applicable to both left- and right-wingers (see Nilsson & Jost, 2020).