Data-Centric Statistical Inference Using R and Tidyverse/

...

/

Comparison with Confidence Intervals

Comparison with Confidence Intervals

Learn about the comparison of hypothesis testing with confidence intervals.

We'll cover the following...

One of the great things about the infer package is that we can jump seamlessly between conducting hypothesis tests and constructing confidence intervals with minimal changes! Recall the code from the previous section that creates the null distribution, which in turn is needed to compute the pp-value:

Press + to interact
null_distribution <- promotions %>%
specify(formula = decision ~ gender, success = "promoted") %>% hypothesize(null = "independence") %>%
generate(reps = 1000, type = "permute") %>%
calculate(stat = "diff in props", order = c("male", "female"))

To create the corresponding bootstrap distribution needed to construct a 95% confidence interval 𝑝𝑚−𝑝𝑓𝑝_𝑚 − 𝑝_𝑓, we only need to make two changes. First, we remove the hypothesize() step because we’re no longer assuming a null hypothesis H0H_0 is true. We can do this by deleting or commenting out the hypothesize() line of code. Second, we switch the type of resampling in the generate() step to bootstrap instead of permute.

Press + to interact
bootstrap_distribution <- promotions %>%
    specify(formula = decision ~ gender, success = "promoted") %>% 
    # Change 1 - Remove hypothesize():
    # hypothesize(null = "independence") %>%
    # Change 2 - Switch type from "permute" to "bootstrap": 
    generate(reps = 1000, type = "bootstrap") %>%
    calculate(stat = "diff in props", order = c("male", "female"))

Using this bootstrap_distribution, let’s first compute the percentile-based confidence intervals:

Press + to interact
percentile_ci <- bootstrap_distribution %>% get_confidence_interval(level = 0.95, type = "percentile")
percentile_ci

Using our shorthand interpretation for 95% confidence intervals, we’re 95% confident that the true difference in population proportions 𝑝𝑚−𝑝𝑓𝑝_𝑚 − 𝑝_𝑓 ...