Abstract
AMOS implements structural modeling. Conceived by James Arbuckle, a psychometrician, as a vehicle for teaching causal analysis, AMOS has evolved into a powerful research environment that eases even nonstatisticians into this important arsenal of tools. AMOS supports structural modeling through maximum likelihood, unweighted least- squares, generalized least squares, Browne's asymptotically distribution-free criterion, and scale-free least-squares. It uses built-in bootstrapping (resampling) to derive standard errors and confidence intervals for sample means, variances, covariances, correlations, and other estimates. Given the researcher's specification of a path diagram, AMOS estimates the path parameters in diagrammatic form, which can be exported to either word processors or graphics programs. The AMOS environment also provides a graphical method for teaching statistics, from hypothesis tests of means through analysis of covariance, multiple regression, and factor analysis, all within the unifying framework of the structural equation modeling perspective.
Get full access to this article
View all access options for this article.
