blinding: Estimating Effect Size using Blinding Survey Data

In bda: Binned Data Analysis

Estimating Effect Size using Blinding Survey Data

Description

To estimate the effect size from an RCT with blinding survey data.

Usage

  blinding.est(x, group, guess, type='cpe')

Arguments

x	outcome variable, continuous, missing values not allowed.
group	arm/group assignment; must have two levels. Missing values not allowed.
guess	Responses from blinding survey, can be two (0=sham/1=active treatment) or three (0=sham/1=active treatment/2=I don't know) levels.
type	three options: "simple" - ignoring blinding survey result. "adjusted" - unblinding adjusted estimate using 'guess' as a covariate in a regression model. "cpe" - change-point approach.

Details

Point estimate, bootstrapping estimate, with/without multiple imputation if missing values exist in 'guess'. TBA

Value

TBA

Examples

 x <- sort(rnorm(20,6,2))
 y1 <- 1 + 1.5 * x[1:10] + rnorm(10)
 y2 <- 10 + 0.5 * x[11:20] + rnorm(10)
 y <- c(y1,y2)
 plot(y~x)
 abline(a=10,b=0.5, col='red',lty=2)
 abline(a=1,b=1.5, col='blue',lty=2)

bda documentation built on Feb. 11, 2026, 9:09 a.m.