ABSTRACT
The Covid-19 pandemic has given further centrality to science within the public debate. But it has also acted as a great multiplier for pseudoscientific (conspiracy) theories. This exploratory study investigates the determinants of pseudoscientific beliefs in five European countries, using data from a survey conducted in May 2021. The concept of pseudoscience is theoretically framed and then operationalised by constructing a Pseudo-scientific Beliefs Index (PBI). Results show that exposure to scientific information does not ‘protect’ against unsound scientific claims, if not complemented by a correct understanding of the division of scientific labour. Pseudoscientific views are strongly associated with distrust of official science. But, in the context of today’s information abundance, even more relevant is the spread of epistemological populism, which fosters reliance on alternative sources and the pseudo-expertise of ‘alternative scientific authorities’. The embrace of ‘alternative scientific facts’ is also associated with electoral support for populist parties.
Acknowledgement
I wish to thank Demos & Pi and Unipolis for having provided the data used in the article. Special thanks go to the journal editors and the anonymous reviewers, whose useful comments have significantly improved the article.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Supplementary material
Supplemental data for this article can be accessed online at https://doi.org/10.1080/14782804.2023.2177838
Notes
1. Considering 2019 OECD data, this index was: Germany 3.190; France 2.196; The Netherlands 2.184; United Kingdom 1.756; Italy 1.466.
2. Parties over 5% were selected using the information provided by Politico’s «Poll of polls» (www.politico.eu/europe-poll-of-polls/) and the Europe Elects project (https://europeelects.eu/).
3. For the same reasons, beta coefficients are also reported for categorical predictors, although standardisation does not permit a substantive interpretation of this parameter.
4. Low and high levels were identified by dividing cases lower than the median value and cases equal to or higher than the median value.
5. The full model specification and results are provided in Table A.10a-e in the online appendix.