The vignette of infer
package gives examples to test equality for ALL categories but not for a SUBSET of the categories.
For example, in infer::gss
dataset, is there a way to test for the income
variable if the proportion of $25000 or more
is equal to 20000-24999
?
Thank you
We can filter
the levels of the 'income', remove the unused levels (droplevels
) and use that in the test
library(dplyr)
library(infer)
gss %>%
filter(income %in% c("$20000 - 24999", "$25000 or more" ) ) %>%
droplevels %>%
specify(response = income, success = "$20000 - 24999") %>%
hypothesize(null = "point", p = .5) %>%
generate(reps = 1000) %>%
calculate(stat = "prop")