twoxtwo

build status (main) build status (dev)

The twoxtwo package provides a collection of functions to display, summarize, and analyze data in two-by-two contingency tables. Statistical analysis functions are oriented towards epidemiological investigation of exposure/outcome relationships.

Installation

## install.packages("devtools")
devtools::install_github("vpnagraj/twoxtwo", build_vignettes = TRUE)

Features

Usage

Example

First load twoxtwo and dplyr to help prep data:

library(twoxtwo)
library(dplyr)

Next create a object with S3 class twoxtwo. For this example, use the twoxtwo::titanic dataset. Note that “exposure” and “outcome” variables must each be binary variables:

crew_2x2 <-
  titanic %>%
  twoxtwo(.data = ., exposure = Crew, outcome = Survived)

crew_2x2
# |         |           |OUTCOME      |OUTCOME     |
# |:--------|:----------|:------------|:-----------|
# |         |           |Survived=Yes |Survived=No |
# |EXPOSURE |Crew=TRUE  |212          |673         |
# |EXPOSURE |Crew=FALSE |499          |817         |

The twoxtwo class has its own summary.twoxtwo() method that computes effect measures (odds ratio, risk ratio, and risk difference):

summary(crew_2x2)
# 
# |         |           |OUTCOME      |OUTCOME     |
# |:--------|:----------|:------------|:-----------|
# |         |           |Survived=Yes |Survived=No |
# |EXPOSURE |Crew=TRUE  |212          |673         |
# |EXPOSURE |Crew=FALSE |499          |817         |
# 
# 
# Outcome: Survived
# Outcome + : Yes
# Outcome - : No
# 
# Exposure: Crew
# Exposure + : TRUE
# Exposure - : FALSE
# 
# Number of missing observations: 0
# 
# Odds Ratio: 0.516 (0.426,0.624)
# Risk Ratio: 0.632 (0.551,0.724)
# Risk Difference: -0.14 (-0.178,-0.101)

Individual measures of effect, hypothesis tests, and impact numbers can be calculated using the twoxtwo object. For example:

crew_2x2 %>%
  odds_ratio()
# # A tibble: 1 x 6
#   measure    estimate ci_lower ci_upper exposure         outcome         
#   <chr>         <dbl>    <dbl>    <dbl> <chr>            <chr>           
# 1 Odds Ratio    0.516    0.426    0.624 Crew::TRUE/FALSE Survived::Yes/No
crew_2x2 %>%
  chisq()
# # A tibble: 1 x 9
#   test      estimate ci_lower ci_upper statistic    df   pvalue exposure outcome
#   <chr>     <lgl>    <lgl>    <lgl>        <dbl> <int>    <dbl> <chr>    <chr>  
# 1 Pearson'… NA       NA       NA            46.5     1 8.97e-12 Crew::T… Surviv…

Note that data analysis can also be performed without first creating the twoxtwo object:

titanic %>%
  odds_ratio(.data = ., exposure = Crew, outcome = Survived)
# # A tibble: 1 x 6
#   measure    estimate ci_lower ci_upper exposure         outcome         
#   <chr>         <dbl>    <dbl>    <dbl> <chr>            <chr>           
# 1 Odds Ratio    0.516    0.426    0.624 Crew::TRUE/FALSE Survived::Yes/No
titanic %>%
  chisq(.data = ., exposure = Crew, outcome = Survived)
# # A tibble: 1 x 9
#   test      estimate ci_lower ci_upper statistic    df   pvalue exposure outcome
#   <chr>     <lgl>    <lgl>    <lgl>        <dbl> <int>    <dbl> <chr>    <chr>  
# 1 Pearson'… NA       NA       NA            46.5     1 8.97e-12 Crew::T… Surviv…

Vignettes

The package includes vignettes to describe usage in more detail.

For details on the twoxtwo data structure and demonstration of basic usage:

vignette("basic-usage", package = "twoxtwo")

For formulas and examples of how to calculate measures of effect:

vignette("measures-of-effect", package = "twoxtwo")

For information on hypothesis testing functionality in the package:

vignette("hypothesis-testing", package = "twoxtwo")

For formulas and demonstration of attributable fraction and impact number calculations:

vignette("af-impact", package = "twoxtwo")

Contributing

Please use GitHub issues to report bugs or request features. Contributions will be reviewed via pull requests.