mvpsmod: Multivariate linear spatial propensity score model.

In czigler/HEIfunctions: Functions for HEI Air pollution accountability grant

Description

Multivariate linear spatial propensity score model. Jointly models the potential outcomes, rather than separate linear models.

Usage

mvpsmod(formula, data, trt, coords, nsamp, nburn, thin, tuning = list(A = 0.1, psi = 0.2, theta = 1), prior = list(KIG = rep(0.01, 2), psi = rep(0.01, 2), theta1 = rep(0.6, 2), theta2 = rep(10, 2)), starting = list(B = rep(0, 2 * length(all.vars(formula))), A = NULL, psi = NULL, theta = NULL))

Arguments

formula	function of propensity score. do not include trt.
data	dataset
trt	string with name of treatment variable; must be defined as 1 if intervention, 0 if no intervention
coords	spatial coordinates corresponding to data.
nsamp	total number of samples
nburn	number of samples to discard as burn in
thin	thinning of samples
tuning	tuning parameters
prior	priors
starting	starting values

Value

samples	MCMC samples for parameters
y0	MCMC samples for y0 (if y0 known, it is constant)
y0	MCMC samples for y1 (if y1 known, it is constant)
coords	spatial coordinates corresponding to data
trt	vector indicating whether intervention or no intervention

Author(s)

Cory

czigler/HEIfunctions documentation built on May 14, 2019, 1:43 p.m.