pathAnal: Perform path analysis at different time lags

Description

Compute the coefficient associated to each path between one or more starting variables and one ending variable at one or more time lags.

Usage

pathAnal(x, from = NULL, to = NULL, lag = NULL, cumul = FALSE, conf = 0.95)

Arguments

An object of class dlsem.

from

The name of the starting variable, or a vector of names of starting variables.

The name of the ending variable.

lag

A non-negative integer or a vector of non-negative integers indicating the time lag at which path coefficients must be computed. If NULL, all relevant time lags are considered.

cumul

Logical. If FALSE (the default), instantaneous path coefficients are returned, otherwise cumulative ones are provided.

conf

The confidence level for edges in the object of class dlsem. Default is 0.95.

Value

Each matrix contains point estimate and asymptotic conf*100% confidence interval for the associated coefficient at each requested time lag.

References

J. Pearl (2012). The Causal Foundations of Structural Equation Modelling. In: R. H. Hoyle (ed.), Handbook of Structural Equation Modelling, Chapter 5. Guilford Press, New York, US-NY.

S. Wright (1934). The Method of Path Coefficients. Annals of Mathematical Statistics, 5(3): 161-215.

Examples

Run this code

data(agres)
mycode <- list(
  GVA~quec(NPATENT,1,15),
  PPI~quec(NPATENT,0,13)+quec(GVA,0,14),
  ENTR_INCOME~quec(NPATENT,0,14)+quec(GVA,1,14)
  )
myfit <- dlsem(mycode,group="COUNTRY",exogenous=c("GDP","FARM_SIZE"),
  data=agres,uniroot.check=TRUE,imputation=FALSE,log=TRUE)
pathAnal(myfit,from="NPATENT",to="PPI",lag=c(0,5,10,15),cumul=TRUE)

Run the code above in your browser using DataLab

Get 50% off unlimited learning