Description Usage Arguments Details Value Author(s) References Examples
Estimates the projections of the identification region along each coefficient dimension for a linear or logistic regression model with interval-censored outcomes (Beresteanu and Molinari 2008, Corollary 4.5). If requested, uses a nonparametric bootstrap to estimate critical values for hypothesis tests about these projections (Beresteanu and Molinari 2008, Algorithm 4.2).
1 2 3 |
formula |
Model formula of the form |
data, subset, na.action |
As in |
model |
|
boot |
Number of bootstrap iterations used to estimate the critical values for inference. |
cluster_id |
Vector of cluster IDs for cluster bootstrap. If
|
maxit |
Maximum number of iterations for the approximation in logistic
regression models. Ignored when |
remove_collinear |
How to treat boostrap iterations in which the design
matrix is rank-deficient. If |
return_boot_est |
Whether to include the bootstrap estimates of the coefficient bounds in the returned object. |
In the linear case, implements largely the same functionality as
oneDproj
and CI1D
in Beresteanu et al.'s (2010) Stata
program.
A list of class "coefbounds"
containing:
coefficients
Matrix containing the sample estimates of the coefficient bounds.
dist
List of matrices containing the bootstrap Hausdorff distances (undirected and directed) used for inference.
boot_est
(if requested) List of matrices of bootstrap estimates of the coefficient bounds.
nobs
Number of observations used in fitting.
call
Original function call.
model
Model used.
Brenton Kenkel
Arie Beresteanu and Francesca Molinari. 2008. "Asymptotic Properties for a Class of Partially Identified Models." Econometrica 76 (4): 763–814.
Arie Beresteanu, Francesca Molinari and Darcy Steeg Morris. 2010. "Asymptotics for Partially Identified Models in Stata." https://molinari.economics.cornell.edu/programs.html
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | ## Simulate data
set.seed(18)
x1 <- rnorm(50)
x2 <- rnorm(50)
y <- 1 - x1 + x2 + rnorm(50)
yl <- floor(y)
yu <- ceiling(y)
## Fit model without covariates
fit_mean <- coefbounds(yl + yu ~ 1, boot = 0)
all.equal(coef(fit_mean)[1, "lower"], mean(yl))
all.equal(coef(fit_mean)[1, "upper"], mean(yu))
## Fit model with covariates
fit_full <- coefbounds(yl + yu ~ x1 + x2, boot = 10)
coef(fit_full)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.