Crosstable

Andrés Necochea

2021-11-15

Crosstable is an extension for the class table that have useful methods to print the table with percents and chisquare expected values. You can use as any regular table call functions like apply, margin.table and chisq.test.

Basic Usage

This function is designed to be used in a way that you can see the results in the console.

library(multitabulation)
#> Loading required package: magrittr
#> Loading required package: stringr
library(magrittr)

# Generate Random data
gender <- sample(c(1,2), 131, replace=TRUE) %>%
  factor(levels=c(1,2), labels=c("Man", "Woman"))
strata <- sample(c(1,2,3), 131, replace=TRUE) %>%
  factor(levels=c(1,2,3), labels=c("Low", "Middle", "High"))
party <- sample(c(1,2), 131, replace=TRUE) %>%
  factor(levels=c(1,2), labels=c("Right", "Left"))

crosstable(strata, party, gender, stats=c("count", "row", "column", "total"))
#> |----------------------------------------------------------------------------------------|
#> |      |     |                                  gender                                   |
#> |      |     |---------------------------------------------------------------------------|
#> |      |     |                 Man                 |                Woman                |
#> |      |     |---------------------------------------------------------------------------|
#> |      |     |              statistic              |              statistic              |
#> |      |     |---------------------------------------------------------------------------|
#> |strata|party|count|% of row|% of column|% of total|count|% of row|% of column|% of total|
#> |------|-----|-----|--------|-----------|----------|-----|--------|-----------|----------|
#> |Low   |Right|    7|  11.67%|     36.84%|     5.34%|   12|  16.90%|     63.16%|     9.16%|
#> |      |Left |    7|  11.67%|     53.85%|     5.34%|    6|   8.45%|     46.15%|     4.58%|
#> |Middle|Right|   15|  25.00%|     51.72%|    11.45%|   14|  19.72%|     48.28%|    10.69%|
#> |      |Left |   11|  18.33%|     39.29%|     8.40%|   17|  23.94%|     60.71%|    12.98%|
#> |High  |Right|    8|  13.33%|     47.06%|     6.11%|    9|  12.68%|     52.94%|     6.87%|
#> |      |Left |   12|  20.00%|     48.00%|     9.16%|   13|  18.31%|     52.00%|     9.92%|
#> |----------------------------------------------------------------------------------------|

Using flextable

You can use as_flextable to make a table ready to be printed in a report. You can combine flextable with officer in order to export the table to Ms Office.

library(multitabulation)
library(flextable)
library(magrittr)

crosstable(strata, party, gender, stats=c("count", "column")) %>%
  as_flextable

Performing Chi Squared Test

Crosstable is basically an extension of table class, so you can use in almost the same way than a regular table, so you can perform the chisq.test

tab <- with(MASS::bacteria, crosstable(y, trt, stats=c("count", "column")))
tabxsq <- chisq.test(tab)
tabxsq
#> 
#>  Pearson's Chi-squared test
#> 
#> data:  tab
#> X-squared = 6.6585, df = 2, p-value = 0.03582

Also, you can include the expected values of chisq.test in the table report using add.tables.

add.tables(tab, format="(0.00)", "(expected)" = tabxsq$expected) %>%
  as_flextable

Aditional options

If you prefer the stats on rows, you can use stats.on.cols=FALSE

crosstable(strata, party, gender, stats=c("count", "column"), stats.on.cols=FALSE) %>%
  as_flextable

You can use an existing table and turn it into a crosstable. You can use col.vars or row.vars

crosstable(Titanic, col.vars=c("Sex", "Survived"), stats=c("count", "column")) %>%
  as_flextable

You can use dnn to set the names of the variables.

with(MASS::bacteria, crosstable(y, trt, stats=c("count", "column"), dnn=c("Result", "Treatment"))) %>%
  as_flextable

If you use a wheighted survey or data base, you can use the formula mode.

crosstable(Freq ~ Type + Cont, stats=c("count", "row"), data=MASS::housing) %>%
  as_flextable