The period 1988-1994 saw latent-variable modeling mature into a cohesive framework that integrates measurement models with covariance-structure inference, placing emphasis on the interpretation of latent constructs and the relationship between observed indicators and latent factors. The era expanded concomitant-variable latent-class models and differential item functioning through log-linear formulations that allow latent-group effects to be examined, while multiway data were addressed through tensorial calibration and direct trilinear decomposition. In predictive settings, methods such as principal component regression, dynamic principal component analysis, and common component analysis were employed to enhance prediction and discrimination across groups.