An R package for tidyverse-friendly statistical inference
The objective of this package is to perform statistical inference using an expressive statistical grammar that coheres with the
tidyversedesign framework. The package is centered around 4 main verbs, supplemented with many utilities to visualize and extract value from their outputs.
specify()allows you to specify the variable, or relationship between variables, that you’re interested in.
hypothesize()allows you to declare the null hypothesis.
generate()allows you to generate data reflecting the null hypothesis.
calculate()allows you to calculate a distribution of statistics from the generated data to form the null distribution.
To learn more about the principles underlying the package design, see
vignette("infer").
<!-- -->
To install the current stable version of
inferfrom CRAN:
install.packages("infer")
To install the developmental stable version of
infer, make sure to install
remotesfirst. The
pkgdownwebsite for this version is at https://infer.tidymodels.org/.
install.packages("remotes") remotes::install_github("tidymodels/infer")
We welcome others helping us make this package as user-friendly and efficient as possible. Please review our contributing and conduct guidelines. By participating in this project you agree to abide by its terms.
For questions and discussions about tidymodels packages, modeling, and machine learning, please post on RStudio Community. If you think you have encountered a bug, please submit an issue. Either way, learn how to create and share a reprex (a minimal, reproducible example), to clearly communicate about your code. Check out further details on contributing guidelines for tidymodels packages and how to get help.
These examples are pulled from the “Full infer Pipeline Examples” vignette, accessible by calling
vignette("observed_stat_examples"). They make use of the
gssdataset supplied by the package, providing a sample of data from the General Social Survey. The data looks like this:
# load in the dataset data(gss)take a glimpse at it
str(gss)
## tibble [500 × 11] (S3: tbl_df/tbl/data.frame)$ year : num [1:500] 2014 1994 1998 1996 1994 ...
$ age : num [1:500] 36 34 24 42 31 32 48 36 30 33 ...
$ sex : Factor w/ 2 levels "male","female": 1 2 1 1 1 2 2 2 2 2 ...
$ college: Factor w/ 2 levels "no degree","degree": 2 1 2 1 2 1 1 2 2 1 ...
$ partyid: Factor w/ 5 levels "dem","ind","rep",..: 2 3 2 2 3 3 1 2 3 1 ...
$ hompop : num [1:500] 3 4 1 4 2 4 2 1 5 2 ...
$ hours : num [1:500] 50 31 40 40 40 53 32 20 40 40 ...
$ income : Ord.factor w/ 12 levels "lt $1000"
As an example, we’ll run an analysis of variance on
ageandpartyid, testing whether the age of a respondent is independent of their political party affiliation.Calculating the observed statistic,
F_hat % specify(age ~ partyid) %>% calculate(stat = "F")Then, generating the null distribution,
null_distn % specify(age ~ partyid) %>% hypothesize(null = "independence") %>% generate(reps = 1000, type = "permute") %>% calculate(stat = "F")Visualizing the observed statistic alongside the null distribution,
visualize(null_distn) + shade_p_value(obs_stat = F_hat, direction = "greater")<!-- -->
Calculating the p-value from the null distribution and observed statistic,
null_distn %>% get_p_value(obs_stat = F_hat, direction = "greater")## # A tibble: 1 x 1 ## p_value ## ## 1 0.055Note that the formula and non-formula interfaces (i.e.
age ~ partyidvs.response = age, explanatory = partyid) work for all implemented inference procedures ininfer. Use whatever is more natural for you. If you will be doing modeling using functions likelm()andglm(), though, we recommend you begin to use the formulay ~ xnotation as soon as possible.Other resources are available in the package vignettes! See
vignette("observed_stat_examples")for more examples like the one above, andvignette("infer")for discussion of the underlying principles of the package design.