---
title: "Creating two-way frequency tables"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Creating two-way frequency tables}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
library(qacBase)
```


The **crosstab** function calculates and prints a two-way frequency table. 

Given a data frame, a row variable, a column variable, and a type (frequencies, cell percents, row percents, or column percents) the function returns a table with

* labeled rows and columns   
* frequencies or percents   
* a level labeled \<NA\> for missing values (if `na.rm = FALSE`)   
* a level labeled Total (unless `total = FALSE`)
* a Chi-square test of independence (if `chisquare = TRUE`)

Tables are printed with 2 decimal places for percents (modifiable using `digits=#`). Variables are coerced to factors if necessary. Adding `plot=TRUE` produces a `ggplot2` graph instead of a table. 

In the examples below, the number of car cylinders (*cyl*) is cross-tabulated with the number of gears (*gear*) for 32 automobiles in the *cars74* data frame.

### Frequencies

By default, the crosstab function reports frequency counts for each 
combination of the two categorical variables. The most common car type has 3 gears and 8 cylinders.

```{r}
crosstab(cars74, cyl, gear)

crosstab(cars74, cyl, gear, plot=TRUE)
```

### Cell percents

Cell percents add up to 100% overall all the cells in the table. 25% of all cars in the data frame have 4 gears and 4 cylinders. 

```{r}
crosstab(cars74, cyl, gear, type="percent")

crosstab(cars74, cyl, gear, type="percent", plot=TRUE)
```


### Row percents

Row percents sum to 100% for each row of the table. 86% of 8 cylinder cars have 3 gears.

```{r}
crosstab(cars74, cyl, gear, type = "rowpercent")

crosstab(cars74, cyl, gear, type = "rowpercent", plot=TRUE)
```

### Column percents

Column percents sum to 100% for each column of the table. Only 7% of 3 gear cars have 4 cylinders.

```{r}
crosstab(cars74, cyl, gear, type = "colpercent")

crosstab(cars74, cyl, gear, type = "colpercent", plot=TRUE)
```

### Chi-square test of independence

You can include a test that the two categorical variables are independent, by adding the option `chisquare = TRUE`.

```{r}
crosstab(cars74, cyl, gear, type = "colpercent", chisquare=TRUE)

crosstab(cars74, cyl, gear, type = "colpercent", plot=TRUE, 
         chisquare = TRUE)
```