ttest_theta: Compute bias adjusted t-statistic for parameter theta

In adzemski/netprobitFE: Network probit with fixed effects

Description

Compute bias adjusted t-statistic for parameter theta

Usage

ttest_theta(fit_ML, gammaS = fit_ML$gammaS_hat,
  gammaR = fit_ML$gammaR_hat, theta = fit_ML$theta_hat,
  rho = fit_ML$rho_hat, theta0 = rep.int(0, length(theta)),
  var_meat_estimator = "clustering", trimming_bound = 0,
  force_robust = FALSE)

Arguments

fit_ML	returned object from ML estimation
gammaS	value used for gammaS (default: estimated value)
gammaR	value used for gammaR (default: estimated value)
theta	value used for theta (default: estimated value)
rho	value used for rho (default: estimated value)
theta0	value of theta under null hypothesis (default: zero vector/test significance)
var_meat_estimator	method to compute "meat" of the variance sandwich * "theorem" implement exactly as in theorem (put hats on everything) * "clustering" clustering based on first-order expansion (will not depend on rho_hat! )
trimming_bound	specify trimming of probabilities
force_robust	don't attempt to use speedglm

Value

list with t statistic and auxilary output

adzemski/netprobitFE documentation built on May 17, 2019, 11:40 a.m.