CUB (version 0.1)

CUSH: Main function for CUSH models

Description

Main function to estimate and validate a CUSH model for ordinal responses, with or without covariates to explain the shelter effect.

Usage

CUSH(ordinal, shelter, m=get('m',envir=.GlobalEnv), X = 0, makeplot = TRUE, summary=TRUE)

Arguments

ordinal

Vector of ordinal responses

shelter

Category corresponding to the shelter choice

m

Number of ordinal categories (if omitted, it will be assigned to the number of categories specified in the global environment)

X

Matrix of selected covariates for explaining the shelter effect. If omitted (default), no covariate is included in the model

makeplot

Logical: if TRUE (default) and if no covariate is included in the model, the algorithm returns a graphical plot comparing fitted probabilities and observed relative frequencies, and a plot of the log-likelihood function at the final estimate, compared with the log-likelihood values of the saturated and the uniform models. If only one explicative dichotomous variable is included in the model, then the function returns a graphical plot comparing the distributions of the responses conditioned to the value of the covariate

summary

Logical: if TRUE (default), summary results of the fitting procedure are displayed on screen

Value

An object of the class "CUSH" is a list containing the following results:

estimates

Maximum likelihood parameters estimates

loglik

Log-likelihood function at the final estimates

varmat

Variance-covariance matrix of final estimates (if X=0, it returns the square of the estimated standard error for the shelter parameter $\delta$)

BIC

BIC index for the estimated model

Details

The estimation procedure is not iterative, so a null result for CUSH$niter is produced. The optimization procedure is run via "optim". If covariates are included, the variance-covariance matrix is computed as the inverse of the returned numerically differentiated Hessian matrix (option: hessian=TRUE as argument for "optim"). If not positive definite, it returns a warning message and produces a matrix with NA entries.

References

Capecchi S. and Piccolo D. (2015). Dealing with heterogeneity/uncertainty in sample survey with ordinal data, IFCS Proceedings, University of Bologna Capecchi S. and Iannario M. (2016). Gini heterogeneity index for detecting uncertainty in ordinal data surveys, Metron - DOI: 10.1007/s40300-016-0088-5

See Also

cushforsim, loglikCUSH

Examples

data(relgoods)
dog<-na.omit(relgoods[,49])
m<-10
shelter<-1
model<-CUSH(dog,shelter=shelter,summary=TRUE)
delta<-model$estimates # ML estimates of delta
maxlik<-model$loglik   # Log-likelihood at ML estimates
sqerrst<-model$varmat    # Squared standard error of delta
BIC<-model$BIC
###############################################
### CUSH model with covariates
music<-relgoods[,47]
shelter<-1
cov<-relgoods[,12]
nona<-na.omit(cbind(music,cov))
ordinal<-nona[,1]
smoking<-nona[,2]
modelcov<-CUSH(ordinal,shelter=shelter,X=smoking,summary=FALSE)
omega<-modelcov$estimates
maxlik<-modelcov$loglik
varmat<-modelcov$varmat
BIC<-modelcov$BIC