Created
April 5, 2022 18:03
-
-
Save bschneidr/d163cd38f2b5a08a3f61df7c12460a5c to your computer and use it in GitHub Desktop.
Estimate the variance of non-response bias estimate, using the survey package
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# Generate example population and sample ---- | |
population <- data.frame( | |
vax_status = sample(x = c(0,1), prob = c(0.25, 0.75), size = 1000, replace = TRUE), | |
response_status = sample(x = c("Respondent", "Nonrespondent"), | |
size = 1000, replace = TRUE, prob = c(0.8, 0.2)) | |
) | |
sample_data <- population[sample(x = 1000,size=150),] | |
# Create a survey design object ---- | |
library(survey) | |
sample_data$pop_size <- 1000 | |
svy_design <- survey::svydesign( | |
data = sample_data, | |
ids = ~ 1, fpc = ~ pop_size | |
) | |
# Estimate means for respondents and for overall population ---- | |
svy_design <- transform( | |
svy_design, | |
is_respondent = ifelse(response_status == "Respondent", 1, 0), | |
any_case = 1, | |
respondent_vax_status = vax_status * is_respondent | |
) | |
estimated_means <- svyratio( | |
numerator = ~ respondent_vax_status + vax_status, | |
denominator = ~ is_respondent + any_case, | |
design = svy_design, covmat = TRUE | |
) | |
domain_means <- coef(estimated_means)[c(1,4)] | |
print(domain_means) | |
# Estimate non-response bias ---- | |
svycontrast( | |
stat = estimated_means, | |
contrast = list('bias_estimate' = c(1,0,0,-1)) | |
) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment