AMOS

AMOS is IBM's graphical package for covariance-based structural equation modelling, used alongside SPSS. Models are drawn as path diagrams — rectangles for observed indicators, ellipses for latent constructs, arrows for hypothesised relationships — and estimated directly from the drawing.

It is the standard tool for confirmatory factor analysis and full structural models when the goal is theory testing: it evaluates how well a hypothesised covariance structure reproduces the observed data, and reports the fit indices reviewers expect (CFI, TLI, RMSEA, SRMR and chi-square over degrees of freedom).

Covariance-based estimation assumes multivariate normality and generally wants a larger sample than PLS-SEM, commonly cited as 200 or more depending on model complexity. Standard practice is a two-step approach: establish an acceptable measurement model through CFA first, then estimate the structural paths, since paths built on a poorly fitting measurement model cannot be interpreted.