Table 4.
Model fit for measurement invariance tests.
Fig 1.
Conceptual diagram of the MIMIC model for AIM–N.
Note. Latent factors (circles) indicated by their items (subset shown for clarity); observed covariates (squares: gender, study level, field, AI usage) predict the latent factors. The dashed arrow illustrates a potential DIF path (covariate → item) representing an intercept shift beyond the latent trait. Diagram is schematic and not to scale; no coefficients are shown here. Abbreviations: AIM–N, MIMIC, DIF.
Fig 2.
MIMIC model results for AIM–N (standardized coefficients).
Note. Standardized path coefficients (β) are displayed only for significant predictors (p <.05); non-significant paths omitted for clarity. Dashed arrow denotes a retained DIF effect (e.g., Field → “stay competitive”/ “degree losing value”). Model fit for the results diagram: χ²(270) = 603.5, CFI =.958, RMSEA =.050; N = 904. Abbreviations: DIF = Differential Item Functioning. See Table 4 for numerical B, SE, β and exact p-values.
Table 1.
Descriptive statistics and reliability for AIM-N subscales (N = 904).
Table 2.
Exploratory factor analysis (EFA) for the identification of the latent structure of AIM-N scale.
Table 3.
Correlations among AIM-N Latent Factors (CFA).
Table 5.
MIMIC model regression paths and DIF Effects.
Table 6.
AIM-N scale items, subscales, and standardized CFA Loadings.