This study examined the psychometric properties of the Intelligenz-Struktur-Test (IST) among candidates of the Indonesian Air Force and Pertahanan University. Data were scored dichotomously (A = 1; others = 0). Analyses included Classical Test Theory (reliability indices), Confirmatory Factor Analysis (CFA), Exploratory Factor Analysis (EFA), and network psychometrics (Graphical Gaussian Models). Measurement invariance across gender and age was evaluated descriptively. Reliability was acceptable (α = .76-.85; ω = .78-.86). CFA indicated the bifactor model fit best (CFI = .93, RMSEA = .05). Network analysis identified numerical reasoning (ZR) parcels as most central (strength up to 2.81). Gender networks were configural and congruent (>.85), with no significant mean differences (p >.05). Age correlated negatively with ZR (r = -.18) and positively with WA (r = .12). IST is a reliable and valid instrument for defense selection in Indonesia. Results highlight both a general factor and domain-specific abilities, with ZR central to the network structure. This study contributes methodologically by integrating latent and network models in a high-stakes context. Keywords: IST, reliability, validity, CFA, network analysis
Copyrights © 2026