The investigators reported divergent factor structures across the two national samples. In the Lebanese cohort, responses aggregated into a largely unitary distress construct. Short sentences capture the core finding: Lebanese adolescents’ answers did not reliably separate into distinct depression, anxiety, and stress scales. Longer, integrative interpretation follows: the authors found that a unidimensional pattern best represented the Lebanese data, which implies that total DASS-Y scores may be the most defensible basis for interpretation in that setting rather than separate subscale scores. By contrast, the Chinese sample displayed clearer differentiation. An ESEM solution better accounted for Chinese response patterns, indicating that depression, anxiety, and stress retained distinctiveness in that group even though a general distress factor remained influential in both samples.
Reliability was consistently high across countries. Subscale composite reliability (CR) ranged from 0.91 to 0.95. Total CR reached 0.97 in both the Lebanese and Chinese samples. The average variance extracted (AVE) for subscales fell between 0.58 and 0.71, with total AVE reported as 0.63 for Lebanon and 0.61 for China. Person separation reliability ranged from 0.79 to 0.85. These numerical indices indicate strong internal consistency and acceptable construct variance capture. The study team emphasized that the structural divergence noted between samples should not be interpreted as a problem of measurement precision, since reliability estimates remained strong throughout.
Rasch analysis identified item-level nonequivalence between the two national groups. Several items functioned non-equivalently, and most of these items belonged to the stress subscale. Short, direct statement: stress items drove much of the cross-national DIF. The study authors therefore cautioned against assuming full item parameter transferability across settings. Longer explanation: if specific items are interpreted or endorsed differently by Lebanese and Chinese adolescents, then cross-national comparisons at the item or subscale level risk conflating true prevalence or severity differences with differences in item functioning. The abstract did not provide the exact item wording or DIF statistics; readers must consult the full article tables and figures for detailed item-level parameters.
The 12-item short form produced psychometric indices comparable to the full 21-item instrument in both samples. Subscale CR for the short form ranged from 0.88 to 0.92, and AVE for subscales ranged from 0.65 to 0.73. Total CR was 0.96 for Lebanon and 0.94 for China; total AVE was 0.64 (Lebanon) and 0.57 (China). These results suggest the abbreviated instrument may offer a practical screening option in school or field settings where time is limited. The study team recommended further cross-national confirmation before the short form is adopted as a universal replacement, given the DIF observed in stress items and the structural differences between national samples.
The analytic design combined multiple modeling strategies. Confirmatory factor analysis, exploratory structural equation modeling (ESEM), and bifactor variants of both were used to test competing dimensional solutions. Rasch analysis evaluated differential item functioning. The samples included 758 Lebanese adolescents (mean age 15.95, SD 1.78; 63.2% female) and 1,062 Chinese adolescents (mean age 16.83, SD 0.59; 47.5% female). Short, factual sentence: reliability and AVE values came from these same analyses. The source stated that there are ethical and legal restrictions on sharing the raw data because participant consent did not permit public resharing; data are available on request from the Jiangxi Psychological Consultant Association for qualified researchers who include institutional affiliation and intended use. The abstract did not report exact item-level parameter estimates, p-values, or model-fit indices; those details appear in the full manuscript’s tables and figures.
Limitations documented in the source were concrete and measurable. The authors indicated that several items showed non-equivalent functioning across countries, and they called for additional cross-national work to confirm the short form and to clarify item-level performance. The abstract did not claim universal measurement invariance; instead, it reported a nuanced outcome: a dominant general distress factor in both samples alongside structural differences in how subdomains map onto responses. This article therefore ends on specific, actionable questions left open by the analyses—further item-level cross-national testing and replication of the short form—rather than broad generalizations about adolescent mental health assessment.