summary.tsci: Summarizing Two Stage Curvature Identification Fits

In TSCI: Tools for Causal Inference with Possibly Invalid Instrumental Variables

Summarizing Two Stage Curvature Identification Fits

Description

Summarizing Two Stage Curvature Identification Fits

Usage

## S3 method for class 'tsci'
summary(object, extended_output = FALSE, ...)

Arguments

object	an object of class 'tsci'.
extended_output	logical. If TRUE are more detailed summary is returned.
...	arguments to be passed to or from other methods.

Value

an object of class 'summary.tsci' containing the following elements:

coefficient

a data frame with columns for the estimated treatment coefficient, its standard error, confidence interval and (two-sided) p-value.

invalidity

a vector containing the number of times the instrumental variable(s) were considered valid, invalid or too weak to perform the test.

viospace_selection

a data frame with columns for the number of times each of the violation space candidate was selected by comparison, the conservative method and as the largest violation space candidate for which the instrumental variable was considered to be strong enough.

treatment_model

a data frame with information about the method used to fit the treatment model.

sample_size_A1

the number of observations in the subset used to fit the outcome model.

sample_size_A2

the number of observations in the subset used to train the parameters for fitting the treatment model.

n_splits

the number of sample splits performed.

mult_split_method

the method used to calculate the standard errors and p-values if n_splits is larger than 1.

alpha

the significance level used.

iv_strength

a data frame with columns containing the estimated instrumental variable strength and the estimated instrumental variable strength threshold for each violation space candidate. Will only be returned if extended_output is true.

coefficients_all

a data frame with columns for the estimated treatment coefficients, its standard errors, confidence intervals and (two-sided) p-values for each violation space candidate.

TSCI documentation built on Oct. 10, 2023, 1:06 a.m.