1. You used Principal Component Analysis (PCA) for factor extraction but then applied Confirmatory Factor Analysis (CFA) to the same dataset without a separate cross-validation sample. PCA is a data reduction technique, not true factor analysis; using it to determine the factor structure and then testing that structure on the identical data guarantees artificially good fit statistics. How can you claim a validated structure when you didn’t test it on a holdout sample or use proper EFA (Common Factor Analysis) from the start?
2. The “Satisfaction” factor showed lower internal reliability, and several key items (e.g., 3.4 on variability, 3.6 on patient harm) had extremely low explained variance (R2<0.14R2<0.14). If core items barely correlate with their intended constructs, isn’t the factorial validity of the HAND-Q fundamentally unstable, particularly regarding how nurses perceive safety versus satisfaction?
3. The entire validation relies on a single-center sample from one University Hospital in Italy, where 61% of nurses have >10 years of experience. Given that handover culture is highly context-dependent, how can you assert this tool is ready for broad evidence-based practice without multi-site discriminant validation to prove it isn’t just measuring local institutional culture?