experiment (version 1.2.1)

ATOPnoassumption: Bounding the ATOP when some of the Outcome Data are Missing Under the Matched-Pairs Design

Description

This function computes the no assumption bounds on the average treatment effect among always-observed pairs (ATOP) when some of the outcome data are missing. The confidence intervals for the ATOP are also computed.

Usage

ATOPnoassumption(Ya, Yb, Ra, Rb, Ta, Tb, l, u, alpha, rep)

Arguments

Ya

A vector of the outcomes of the first unit in the matched pairs. The missing values for Ya should be coded as NA.

Yb

A vector of the outcomes of the second unit in the matched pairs. The missing values for Yb should be coded as NA.

Ra

A vector of the missing data indicators of the first unit in the matched pairs.

Rb

A vector of the missing data indicators of the second unit in the matched pairs.

Ta

A vector of the treatment conditions of the first unit in the matched pairs.

Tb

A vector of the treatment conditions of the second unit in the matched pairs.

l

The lower limit of the outcome.

u

The upper limit of the outcome.

alpha

A positive scalar that is less than or equal to 0.5. This will determine the (1-alpha) level of confidence intervals. The default is 0.05.

rep

The number of repetitions for bootstraping.

Value

A list of class ATOPnoassumption which contains the following items:

LB

The lower bound for the ATOP.

UB

The upper bound for the ATOP.

LB.CI

The lower limit of the confidence interval for the ATOP.

UB.CI

The upper limit of the confidence interval for the ATOP.

Details

For the details of the method implemented by this function, see the references.

References

Kosuke Imai and Zhichao Jiang (2018). “A Sensitivity Analysis for Missing Outcomes Due to Truncation-by-Death under the Matched-Pairs Design”, Technical Report. Department of Politics, Princeton University.

Examples

# NOT RUN {
data(seguro)
attach(seguro)
ATOPnoassumption(Ya,Yb,Ra,Rb,Ta,Tb,l=0,u=1,alpha=0.05,rep=100)
# }