14-Item Scale Tracks Related Moral and Civic Identities in German-Speaking Youth

A 14-item questionnaire developed with 1,149 adolescents and emerging adults in Germany and German-speaking Switzerland found that moral identity and civic identity are strongly related but can be measured as empirically distinguishable components of the self-concept. The Moral–Civic Identity Scale, or MDI, asks how important moral traits and civic or democratic principles are to each respondent’s identity. In the final two-factor model, the two domains were highly correlated, at r = 0.780, but a correlated two-factor solution fitted the data better than a single general factor. The findings provide initial, culturally bounded validation of a tool designed to assess identity rather than behavior: how central qualities such as honesty and compassion, or principles such as equality and tolerance, are to a person’s sense of self.

The study was cross-sectional and designed primarily to develop and validate the scale, not to establish developmental changes or causal effects. Participants aged 16 to 27 were recruited through a professional online panel, with quotas intended to balance gender and approximate a residence distribution of 70% Germany and 30% German-speaking Switzerland. The final sample included 807 residents of Germany and 342 residents of the German-speaking part of Switzerland. Its mean age was 22.98 years, although 42.9% were aged 25 to 27 and only 12.3% were 16 to 18. Because it was a nonprobability panel sample, the results should not be read as nationally representative estimates of young people’s identities or attitudes.

Researchers extended an established German Moral Identity Scale with nine newly written civic traits: politically engaged, tolerant, justice-oriented, equality-oriented, gender-equitable, diversity-conscious, solidary, critically reflective and civically courageous. The civic pool was derived through a theory-driven review of literature on civic and democratic development, including the Council of Europe’s Reference Framework of Competences for Democratic Culture and the International Civic and Citizenship Education Study. The team then reviewed the proposed descriptors for conceptual fit, content coverage, clarity and compatibility with the moral scale’s single-trait format. Respondents rated all 27 traits, including nine distractors, on a four-point scale ranging from “not at all important to me” to “very important to me.” The distractors were not included in the statistical analyses.

Four prespecified models were compared with confirmatory factor analysis: a single general factor, a correlated two-factor model separating moral from civic identity, a higher-order model and a bifactor model combining a general factor with domain-specific factors. The one-factor model fitted poorly, with a robust comparative fit index of 0.826, a root mean square error of approximation of 0.081 and a standardized root mean square residual of 0.064. The correlated two-factor and higher-order specifications improved fit, but the bifactor model produced the strongest global statistics, including a robust CFI of 0.929. Researchers nevertheless rejected it as the preferred structure because several specific-factor loadings were very low or negative, cross-loadings were meaningful, and the general factor did not cleanly dominate the item variance. Although it was statistically competitive, the bifactor solution was less theoretically coherent and interpretable than the simpler correlated two-factor model.

Because the best-fitting bifactor structure remained difficult to interpret, investigators conducted post hoc item-reduction analyses within the two-factor solution. They removed “politically engaged” and “selfless” in a first stage, then “critically reflective” and “generous” after re-estimating the model. Decisions combined weak standardized loadings, low explained variance, local misfit and consideration of content coverage. The resulting 14-item model showed improved fit, with a robust CFI of 0.913, RMSEA of 0.071 and SRMR of 0.054. All retained items had statistically significant standardized loadings from 0.49 to 0.71, and item-level R² values ranged from 0.24 to 0.51. The authors describe the fit as acceptable rather than excellent, making the shortened model a candidate structure rather than a fully independently confirmed instrument.

The two subscales had good internal consistency. Moral identity produced a Cronbach’s alpha of 0.80 and omega of 0.81, while civic identity produced alpha and omega values of 0.82. Composite reliability was 0.80 for moral identity and 0.82 for civic identity. Their external associations were nevertheless modest and overlapped, which the researchers interpret as initial nomological evidence rather than definitive proof of discriminant or incremental validity. Moral identity was more strongly associated than civic identity with lower violence support and lower aggression. The respective moral-identity correlations were r = −0.16 and r = −0.19, compared with r = −0.06 and r = −0.07 for civic identity; formal tests confirmed the differences between these dependent correlations.

The pattern reversed for civic-democratic attitudes. Civic identity was more strongly negatively associated with political deprivation and distance from democratic principles than moral identity was. Its correlations with those measures were r = −0.17 and r = −0.32, while moral identity correlated at r = −0.10 and r = −0.24. Statistical comparisons supported the stronger civic associations. All four external criteria, however, were related to both identity domains, and the identity factors themselves were strongly correlated. The results therefore suggest that measuring the two domains separately can reveal differences in their patterns of association, not that moral and civic identity operate independently. The study also tested whether the shortened scale behaved comparably across the two countries and between male and female participants.

Multigroup analyses supported full measurement invariance across residence: the structure, loadings, intercepts and residual variances were comparable between Germany and German-speaking Switzerland. For gender, the analysis supported configural, metric and scalar invariance, but not strict invariance, indicating that residual variances may differ between male and female participants. The researchers therefore described gender comparability as partial: latent means and structural relations can be compared cautiously, while item-level measurement errors require care. Only 10 participants identified as gender diverse, so the gender analysis excluded that group and cannot establish measurement equivalence across gender diversity. The scale is currently intended for use in the two German-speaking settings studied, not as a generally validated international instrument.

Several limitations constrain the findings. The final items were selected using the same sample used to compare the original models, meaning that independent cross-validation is still required. The cross-sectional design cannot show whether moral and civic identities become more differentiated over time, whether civic opportunities shape civic identity, or whether changes in one domain are prospectively associated with changes in the other. Self-report ratings may also reflect social desirability and a general tendency to endorse positively valued traits, while the distractor items were not designed to measure those response styles directly. Age showed no significant relationship with moral identity, although it had a weak positive association with civic identity at r = 0.07. Educational attainment was likewise unrelated to moral identity but weakly associated with civic identity at Spearman’s ρ = 0.18. These cross-sectional patterns are descriptive, not evidence that age or schooling causes identity development.

Overall, the MDI offers researchers a short, openly available instrument for asking whether moral and democratic principles are central to a young person’s self-definition. Its strongest contribution is practical rather than dramatic: it puts interpersonal moral traits and civic-democratic values into a parallel format, allowing investigators to measure their substantial overlap while testing whether each domain has a distinct pattern of external associations. The scale may support research and program evaluation concerning moral and civic development, but broader claims await probability-based samples, independent replication, longitudinal designs, behavioral measures and cross-cultural validation. For now, the evidence supports the 14-item MDI as a promising initial tool for German-speaking adolescents and emerging adults, with the clearest distinctions lying in its differential associations and the boundaries of its intended cultural setting.

Subject of Research: Development and initial validation of a moral–civic identity scale for adolescents and emerging adults in Germany and German-speaking Switzerland

Article Title: Scale development and factor analysis of the moral–civic identity scale in adolescents and emerging adults from Germany and German-speaking Switzerland

Article References: Wilke, J., von Düring, U., & Zumbach-Basu, J. (2026). Scale development and factor analysis of the moral–civic identity scale in adolescents and emerging adults from Germany and German-speaking Switzerland. Discover Psychology, 6(1), Article 290. https://doi.org/10.1007/s44202-026-00951-w

Image Credits: AI Generated

DOI: 10.1007/s44202-026-00951-w

Keywords: moral identity, civic identity, scale development, psychometrics, adolescents, emerging adulthood, confirmatory factor analysis, measurement invariance, Germany, German-speaking Switzerland, Scale, development