MLE_stage2: ML estimation stage 2
In adzemski/netprobitFE: Network probit with fixed effects

ML estimation stage 2

1
2
3

MLE_stage2(fit_MLE_stage1, gammaS = fit_MLE_stage1$gammaS_hat,
  gammaR = fit_MLE_stage1$gammaR_hat, method = "mle", rho_start = 0,
  kappa = 0.02)

`fit_MLE_stage1`	returned list from stage 1
`gammaS`	sender effects
`gammaR`	receiver effects
`method`	maximize log likelihood ("mle") or search for root of gradient ("gradient")
`rho_start`	starting value of rho (reciprocity parameter)
`kappa`	bound rho away from -1 and 1 (see paper)