Cultural Continuity and Future Competencies in Aggregated Stated Priorities for Preparing Young People in Kazakhstan

This study examines aggregated stated priorities for preparing young people in a territorially stratified, quota-controlled achieved sample of adults surveyed in Kazakhstan in 2025 (n = 3,123). Respondents selected up to five of 15 qualities and educational goals. The findings measure relative salience within the supplied list; they do not measure perceived social norms, absolute endorsement, or attained competencies. The analysis combines item-level frequencies, bootstrap rank intervals, four domain-aggregation metrics, seven classification schemes, two permutation null models, and respondent-stratified conditional logistic regression. Sensitivity analyses address classification, complete-case selection, inference with 20 regional clusters, and the ordinal coding of material hardship. Qualitative contextualization draws on 17 unique focus group transcripts, 20 expert interviews, and 100 open-ended expert survey responses. Practical responsibility ranked first in mean item-selection rate under all seven classification schemes. Cultural-continuity and civic-engagement estimates were sensitive to domain width and item assignment. Co-selection conclusions depended on the null model: no pair exceeded the primary item-margin expectation, whereas the category-margin model identified positive departures for selected pairs. Adding 27 sociodemographic interactions yielded only a small improvement over item fixed effects (incremental McFadden R² = .0057). Russian-language self-identification was associated with a greater relative allocation to adaptive-cognitive items, although statistical support depended on the inference procedure. The material-hardship association with civic allocation did not survive Holm adjustment in the small-cluster score-bootstrap analysis, and the categorical specification did not support a uniform gradient. Policy interpretation of these constrained selections requires item-level transparency, normalization for domain width, and explicit specification of classification and inferential assumptions.This study examines aggregated stated priorities for preparing young people in a territorially stratified, quota-controlled achieved sample of adults surveyed in Kazakhstan in 2025 (n = 3,123). Respondents selected up to five of 15 qualities and educational goals. The findings measure relative salience within the supplied list; they do not measure perceived social norms, absolute endorsement, or attained competencies. The analysis combines item-level frequencies, bootstrap rank intervals, four domain-aggregation metrics, seven classification schemes, two permutation null models, and respondent-stratified conditional logistic regression. Sensitivity analyses address classification, complete-case selection, inference with 20 regional clusters, and the ordinal coding of material hardship. Qualitative contextualization draws on 17 unique focus group transcripts, 20 expert interviews, and 100 open-ended expert survey responses. Practical responsibility ranked first in mean item-selection rate under all seven classification schemes. Cultural-continuity and civic-engagement estimates were sensitive to domain width and item assignment. Co-selection conclusions depended on the null model: no pair exceeded the primary item-margin expectation, whereas the category-margin model identified positive departures for selected pairs. Adding 27 sociodemographic interactions yielded only a small improvement over item fixed effects (incremental McFadden R² = .0057). Russian-language self-identification was associated with a greater relative allocation to adaptive-cognitive items, although statistical support depended on the inference procedure. The material-hardship association with civic allocation did not survive Holm adjustment in the small-cluster score-bootstrap analysis, and the categorical specification did not support a uniform gradient. Policy interpretation of these constrained selections requires item-level transparency, normalization for domain width, and explicit specification of classification and inferential assumptions. Received: 10 August 2026 / Accepted: 19 October 2026 / Published: November 2026