Year: 2026
Type: article
Abstract: In this brief report, we show that when performing structural equation modeling, substantive conclusions may differ drastically depending on the statistical program that is used. This can occur because some programs assume that certain model errors are correlated by default, whereas others do not. We argue that the uncritical use of statistical software can lead to situations in which incorrect inferences are made.
Source: Communication Research Reports
Authors: Brian Manata, Franklin J. Boster
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7140029966