FULL PDF
Abstract
Linear structural equation modeling has become an essential methodology for specifying, estimating, and testing hypothesized interrelationships among substantively meaningful variables. Historically, researchers relied on an arcane matrix language to construct and describe models in publications, which, while technically accurate, obscured basic principles that could otherwise be understood by anyone with a reasonable scientific background. This complexity suggested the need for specialized procedures for non-standard applications. Fortunately, many programs in the field have abandoned such approaches, making structural modeling more accessible, though sometimes at the cost of concealing the underlying principles. Since its initial release via BMDP Statistical Software (Bentler, 1985), EQS has revolutionized structural modeling by simplifying the process while maintaining technical sophistication and accuracy. This legacy continues with the release of EQS 6, which expands the range of models, data, and statistical methods available for analysis, delivering simplicity alongside the highest statistical precision. The EQS program manual provides comprehensive instructions for setting up models, running them under various options, and interpreting output across platforms such as mainframes, Unix, Linux, Windows, and, previously, Macintosh computers. While Windows and Mac versions include the text interface described in the manual, they also offer enhanced graphical interfaces and a broader selection of general-purpose statistical methods and plots useful for research. Although only the Windows version is currently available, its features are thoroughly detailed in the EQS 6 for Windows User’s Guide (Bentler & Wu, 2002), continuing the tradition of combining usability with technical rigor to meet researchers' diverse needs.
View more >>