Statistical differences between groups

[Sjoukje Goldman]

This is probably a question to which I could have found an answer, I just seem not to be able to find it or understand what I find. 

I have conducted a SEM multiple group analysis. Now I would like to test if there are significant differences between the two groups per parameter. 

To be more clear:

H1 A -> D

H2 B -> D

H3 C -> D

Now I would like to know if the standardized parameter of groups 1 for A -> D differs significantly from group 2. I think I need to work with constraints, but don't fully understand how to use it. Would anyone be willing to explain this to me? Or give me a hint how to look for the answer? 

Thank you so much!

[Terrence Jorgensen]

Now I would like to know if the standardized parameter of groups 1 for A -> D differs significantly from group 2. 

model <- '
  x4 ~ c(a1, a2)*x1 + c(b1, b2)*x2 + c(c1, c2)*x3
  diff.a := a1 - a2
  diff.b := b1 - b2
  diff.c := c1 - c2
'
fit <- sem(model, data = HolzingerSwineford1939, group = "school")

## standardized differences (:=) are differences between standardized slopes
parameterEstimates(fit, standardized = TRUE)

## test standardized versions of user-defined differences
standardizedSolution(fit)

Terrence D. Jorgensen

Postdoctoral Researcher, Methods and Statistics

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

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