Austrian researcher Andreas Scheu urges caution against ranking countries by science trust

Scheu is affiliated with the University of Klagenfurt, which along with other partners has been examining the ways publics relate to scientific institutions. His comments come amid recurring references in European debate to trust data such as those produced by the EU Eurobarometer, a long-running series of cross-country opinion surveys on topics including science, politics and the economy.

The crux of Scheu's argument is methodological. High aggregate scores on trust in science, he said, are not automatically a sign that a society enjoys particularly favorable conditions for research and evidence-based policy. In some contexts, such numbers may even coexist with comparatively low trust in political institutions, which complicates any simple reading.

What the data actually measure

He therefore urged readers, journalists and policymakers to treat country-level trust figures as one input among several rather than as a league table. By his account, the more useful questions concern which institutions are trusted, by whom and in what circumstances, rather than which nation comes out on top.

Scheu's position has support in his institutional environment. The Berlin-Brandenburg Academy of Sciences and Humanities (BBAW), where he is also active, has long hosted projects on the philosophy and sociology of science. Researchers there have repeatedly cautioned against reducing the messy reality of public attitudes toward expertise to a single percentage point.

Christian Schuster, also associated with the same research milieu, has likewise emphasized that trust in science is shaped by local political cultures, media environments and recent controversies. Such factors, in his view, mean that seemingly identical numbers in two countries can reflect very different social realities.

Permanent self-correction as a strength

The Austrian Academy of Sciences (ÖAW) has likewise documented shifts in public attitudes toward science, including during the recent pandemic period, and has stressed that trust is volatile. Even where overall trust scores are high, the academy has noted, specific groups within a population may express deep skepticism about particular fields such as vaccines, climate modeling or artificial intelligence.

Scheu's broader research line, which he and colleagues sometimes describe with the shorthand Reflektierte — a label that recurs in the source materials — is concerned with how scientific communities understand themselves and their relationship with the wider public. He and his collaborators frame the self-critical capacity of science as one of its defining strengths.

The notion of "permanente Selbstkorrektur," or permanent self-correction, is central to how Scheu presents scientific practice. Science, on this view, is not defined by the certainty of any individual result but by the willingness of its practitioners to revise claims when evidence changes. Trust, he suggests, flows from that visible process rather than from authoritative pronouncing.

Implications for policy

That framing carries policy implications. If high trust figures are read uncritically, governments may relax the work of explaining scientific reasoning to the public. If low trust figures trigger alarm, decision makers may reach for simple reassurance campaigns rather than engaging with substantive doubts. Scheu's caution applies to both temptations.

Speaking to APA, Scheu did not single out any specific country for praise or blame. Instead, he returned repeatedly to the idea that cross-country comparisons can flatten important differences and create an illusion of measurement precision that the underlying survey instruments do not in fact deliver.

The exchange has been picked up by academic commentators in Austria and Germany, several of whom have echoed Scheu's warning in their own commentary. Their joint message is that public trust in science is real and measurable but also contingent, context-dependent and shaped by factors that a single ranking cannot capture.