DE.crit.path: Find critical paths in a Path Analysis model.
In dustinfife/DE: Performs an analysis of Data Elasticity

Description Usage Arguments Details Value Author(s) Examples

This function randomly generates models and determines which paths, when removed, tend to reduce the fit of the model substantially.

1
2
3

  DE.crit.path(data, variable.names = names(data), paths,
    prop.arrows = 0.2, iterations = 50,
    restrictions = NULL, cutpoint)

`data`	The correlation matrix to be fit.
`variable.names`	A list of variables names if they're not specified in the correlation matrix.
`paths`	How many paths should be generated? Can be specified as a single value or as a range. See examples.
`prop.arrows`	What proportion of arrows should be correlational?
`iterations`	Number of random models to be generated.
`restrictions`	A matrix containing restrictions on the DE procedure. See details section.
`cutpoint`	An RMSEA value that differentiates between the modes of the two distributions.

Fife, Rodgers, and Mendoza (2013) noted that DE distributions are frequently bimodal and commented that a possible reason for this is that the poor fitting distribution has one or more “critical paths” that are missing. This function first generates iterations random models, then uses the value for cutpoint to identify which paths exist in the right versus left side of the cutpoint.

returns a matrix the specifies which paths are contained in the "highMode" versus "lowMode" datasets

Dustin Fife

restrictions = matrix(c("", "Age", 0), nrow=1, byrow=TRUE)
data(albanese)

crit = DE.crit.path(data=albanese, paths=c(5,7), prop.arrows=0, iterations = 20, restrictions=restrictions, cutpoint=.4)
crit