keywords Mediation, mediated moderation, moderated mediation, conditional mediation, bootstrap, Sobel test
All the coefficients in the Mediation tables are
estimated using Maximum Likelihood method implemented in lavaan R package. The betas
are obtained as standardized parameters of the path model. They are
often called completely standardized effect size indexes . For
the indirect effect, the beta is also called index of mediation
(Preacher
& Kelley, 2011).
If Individual regression option is selected, the output
shows the individual regressions necessary to estimate the mediation
model. They show the OLS regression coefficients, usually employed in
regression analysis (Baron
& Kenny, 1986)
Due to the different methods of estimation, the estimates obtained in
the Mediation tables and in the
Individual regressions tables may slightly vary, although
they never differ in a substantial way. The maximum likelihood estimates
are more than fine (Rijnhart,
Twisk, Chinapaw, de Boer, Heymans, 2017) The purpose of the
individual regressions tables is to allow exact comparison with results
obtained in R (or other software) using the Baron and Kenny
approach.
Got comments, issues or spotted a bug? Please open an issue on GAMLj at github“ or send me an email