smallarea (version 0.1)

smallareafit: Fits a Fay Herriot Model to data

Description

Fits a Fay Herriot model to the data and returns a list of items which area estimates of different paramaters and mse of the estimates of the small area means the details of which are provided in the value section.

Usage

smallareafit(formula, data, method)

Arguments

formula

an object of class formula. a formula similar in appearance to that of in lm function in R. It has to be ascertained that the data contains a column of the sampling variances, and that while specifying the formula the the name of the variable that contains the sampling variances should preceede the variables which are the covariates. e.g response~D+x1+x2 is a correct way of specifying the formula where as response~x1+D+x2 is not.(note D is the variabe that contains the values of sampling variances and x1 and x2 are covariates). In general the first of the variables on the right hand side of ~ will be treated as the vector of sampling variance. Note that our function does not accept any missing values.

data

an optional data.frame. containg the variable names and data. in absence of this arguments the function will accept the corresponding things from the global environment.

method

method can be from "PR", "FH", "ML", "REML"

Value

smallmean.est

Estimates of the small area mean

smallmean.mse

Mean Square Prediction error of the estimates of the small area mead

var.comp

an estimate of the variance components

est.coef

an estimate of the regression coefficients

Details

for more details see the vignette

References

On measuring the variability of small area estimators under a basic area level model. Datta, Rao, Smith. Biometrika(2005),92, 1,pp. 183-196 Large Sample Techniques for Statistics, Springer Texts in Statistics. Jiming Jiang. Chapters - 4,12 and 13. Small Area Estimation, JNK Rao, Wiley 2003 Variance Components, Wiley Series in Probability and Statistics,2006 Searle, Casella, Mc-Culloh

See Also

prasadraoest fayherriot resimaxilikelihood maximlikelihood

Examples

data=data.frame(response=c(1,2,3,4,8),D=c(0.2,0.5,0.1,0.9,1))
data
ans=smallareafit(response~D,data,method="FH")
ans1=smallareafit(response~D,data,method="REML")
ans2=smallareafit(response~D,data,method="PR")
ans3=smallareafit(response~D,data,method="ML")