[R] Setting up hypothesis tests with the infer library?

[Neal Fultz]

>
> I've been setting up problems like this with code similar to:
> ===========================
> df <- data.frame(
>     survey = c(rep("1980", 1000), rep("2010", 1000)),
>     DP = c(rep("Y", 0.66*1000), rep("N", 1000 - (0.66*1000)),
>            rep("Y", 0.64*1000), rep("N", 1000 - (0.64*1000))))
>


I think infer needs the data in 'long' format. Here is how I would approach
it:

df <- expand.grid(year=c(1980, 2010), DP=c("Approve", "Disapprove"))

expand.grid creates all combos of year and view

df <- df[ rep(1:4, c(660, 640, 1000-660, 1000-640)),  ]

here I used dataframe indexing to repeat each of the four rows. I find this
a lot easier than manually creating columns the way you did.
tidyverse-style packages don't encourage this style though. Tastes will
vary.

In dplyr style, add a column and use it in slice:

df <- expand.grid(year=c(1980, 2010), DP=c("Approve", "Disapprove")) %>%
    mutate(freq = c(660, 640, 1000-660, 1000-640)) %>%
    slice(rep(1:n(), freq))


xtabs(~year+DP, df)

checks that it worked correctly. Or use dplyr::count() if that'swhat you're
comfortable with.

	[[alternative HTML version deleted]]