Selection, Socialisation, and Political Change: A Longitudinal Analysis of Political Authoritarianism in Poland, 1998–2023

How do authoritarian attitudes form and evolve over the adult life course? The literature on authoritarianism has long debated whether individual differences reflect early-life selection—stable dispositions shaped by childhood socialisation and social position—or ongoing socialisation through which attitudes shift in response to life experiences and political contexts. Adjudicating between these mechanisms has proven empirically difficult, as most studies rely on cross-sectional data that cannot distinguish compositional differences between individuals from genuine within-person change. This study addresses this limitation using the Polish Panel Survey (POLPAN), a nationally representative longitudinal study spanning eight waves from 1988 to 2023 at five-year intervals. We analyse six waves from 1998 to 2023 to examine political authoritarianism in Poland—a particularly instructive case given the country's profound political transformations over this quarter-century, including EU accession, democratic consolidation, and a marked authoritarian turn under Law and Justice (PiS) from 2015. These dynamics raise a fundamental question: do citizens' authoritarian attitudes respond to macro-political shifts, or do aggregate changes primarily reflect cohort replacement and differential attrition? We operationalise political authoritarianism as a latent construct measured by three items capturing preferences for concentrated, unconstrained executive power: endorsement of single-party dominance, dismissal of electoral accountability, and rejection of legal constraints. To ensure valid cross-temporal comparison, we conduct longitudinal confirmatory factor analysis establishing partial scalar measurement invariance across waves, extracting factor scores for subsequent panel regression models. Our analytical strategy employs Correlated Random Effects (CRE-Mundlak) models that decompose predictor effects—education, social class, and class mobility—into between-person and within-person components, enabling direct estimation of selection effects versus socialisation effects. We additionally implement Chow tests and time-predictor interactions to assess structural breaks, attending particularly to 2008 and 2013—politically significant junctures preceding PiS's ascent. Three principal findings emerge. First, education and social class influence authoritarianism predominantly through selection rather than socialisation: individuals with higher education and professional-class positions hold less authoritarian attitudes, but these differences are established prior to observation and remain largely stable. Within-person change attributable to educational attainment or class mobility is modest. Second, a statistically significant structural break emerges around 2013, after which relationships between certain predictors and authoritarianism shift—suggesting Poland's changing political context altered attitude-formation dynamics. Third, despite aggregate-level polarisation, individual-level authoritarian attitudes display notable stability, implying that societal shifts owe more to compositional change than mass attitudinal conversion. These findings carry implications for theory and methodology in political psychology. Substantively, they suggest that interventions targeting adult socialisation may have limited efficacy in reducing authoritarian predispositions; selection processes operating earlier in life appear more consequential. Methodologically, the study demonstrates the value of within-between decomposition for disentangling mechanisms that cross-sectional designs conflate—an approach broadly applicable to longitudinal attitude research beyond the Polish case.