Questionable prospective effects between shame, loneliness, and non-suicidal self-injury among Chinese adolescents: A reassessment of Chen et al. (2026)

Chen et al. argued that shame and loneliness play a role in the longitudinal emergence of non-suicidal self-injury among adolescents. However, significant cross-lagged parameters do not by themselves establish genuine prospective influence because similar estimates may arise through alternative statistical processes. Using simulated data that reproduced the sample size and correlation matrix reported by Chen et al., we examined whether their conclusions were robust across several defensible analytic specifications. The pattern of results depended strongly on analytic choice: some models suggested positive prospective associations, whereas others indicated negative or null effects. When estimates from all specifications were combined using meta-analytic methods, no overall effect reached statistical significance. Accordingly, the available evidence does not provide a firm basis for concluding that shame, loneliness, and non-suicidal self-injury exert reciprocal prospective influences on one another. In line with multiverse-analysis recommendations, conclusions should be drawn from the collective pattern across reasonable analytical alternatives rather than from a single model.