Trust in official statistics and quality of governance strengthen trust in institutions and politics. Evidence from OECD survey on trust
To presented at the ISI/IAOS congress in Vilnius May 13 2026
Serge Allegrezza[1]
Wolfgang Langer[2]
Majlinda Joxhe[3]
[1] STATEC, emeritus Director of the National Statistical Institute of Luxembourg, President of STATEC Research asbl, e-mail: serge.allegrezza@ext.statec.etat.lu
[2] Martin-Luther-Universität Halle-Wittenberg, Institut für Soziologie, Associate Professor, e-mail: wolfgang.langer@soziologie.uni-halle.de
[3] STATEC, Researcher for the National Statistical Institute of Luxembourg, e-mail: majlinda.joxhe@ext.statec.etat.lu
Abstract
This study examines the determinants of trust in politics and trust in institutions using data from the OECD 2023 survey “Building trust to reinforce democracy”. In the first step, we conduct a confirmatory factor analysis to validate the latent constructs of trust in politics, trust in institutions, trust in statistics, and perceived quality of governance. The confirmatory model demonstrates strong factor loadings, providing a robust measurement framework for subsequent analysis. Building on this foundation, we use factor scores in multilevel regression models to assess how individual- and country-level determinants shape political and institutional trust. Trust in official statistics consistently emerges as one important predictor, ceteris paribus. Individuals who perceive public statistics as trustworthy, accessible and useful are more likely to express higher trust in institutions and politics. At macrolevel, Institutional trust is determined by GDP, political stability, and corruption intensity. These findings suggests that trust in statistics deserves to be heeded by safeguarding independence and transparency of the statistical system. But objective quality of products and NSIs are not reflected in trust by citizens in statistics. The paper pledges for taking into account the perceptions of the quality of statistics by measuring it systematically and using techniques of inoculation, debunking and fact checking to close the gap between perception and objective quality of statistics.
The powerpoint can be downloaded here.