R-Square Mediation

[Daniel Groß]

Hi all,

I calculated a simple mediation model with lavaan using the following code:

NA.    <- 'ZMHAS_12 ~ a*ZNA
          ZPMH_scale ~ b*ZMHAS_12 + c*ZNA


          indirect := a*b
          direct := c
          total := c + (a*b)'


fit <- sem(NA., data = data, estimator = "MLR")
summary(fit, standardized=T, fit.measures=T, rsq = T)

rsq = T gives me the R-square. Now my question is where I can see whether the R-squares are significant?

Thanks for your answers.

Daniel 

[Terrence Jorgensen]

Now my question is where I can see whether the R-squares are significant?

The H0 that Rsq == 0 is the same as saying all its predictors have slopes == 0.  You can use a Wald test:

# test Rsq for mediator, same as Wald z from summary() but squared
lavTestWald(fit, constraints = 'a == 0')
# test Rsq for outcome
lavTestWald(fit, constraints = 'b == 0 ; c == 0')

Terrence D. Jorgensen

Assistant Professor, Methods and Statistics

Research Institute for Child Development and Education, the University of Amsterdam

http://www.uva.nl/profile/t.d.jorgensen

[Daniel Groß]

Thanks for the good and fast reply.