tabulate
Counts how many rows fall in each category, or in each pair of categories.
What we have
- one-way and two-way tables
row,column,chi2,nofreqandnolabelsummarize(), for the mean of another variable in each cell- frequency weights
Coming later
tab1andtab2, for several tables in one command- the chi-squared statistic is not printed while
set mincellis on, because it can be worked backwards to the hidden counts
You may shorten it to tab.
How you write it
tabulate var1 [var2] [if] [weight] [, row column nofreq chi2]
Square brackets mean optional. A vertical bar means choose one.
What it does
One variable gives a frequency table: each value, how many rows have it, and what share of the whole that is. Two variables give a cross-tabulation — a grid of counts — with a test of whether the two are related (a chi-squared test) if you ask for it.
row and column turn the counts into percentages across or down; nofreq drops the counts and leaves only the percentages.
Watch out. With
set mincell switched on, cells with very few people are hidden, and the chi-squared statistic is not printed at all — because it can be worked backwards to recover the hidden counts. See set mincell.Try it
Every example below was run after sysuse vira.sample02, clear.
. tabulate sector
sector | Freq. Percent Cum.
-------------+------------------------------------
Rural | 3,987 66.45 66.45
Urban | 2,013 33.55 100.00
-------------+------------------------------------
Total | 6,000 100.00
The same thing in pandas and in R
pandas
df["sector"].value_counts().sort_index()
R
library(dplyr) df %>% count(sector) %>% mutate(percent = 100 * n / sum(n)) # This R was run in a real R (4.6.1) and gave the same numbers.
. tabulate state sector, row
sector
state | Rural Urban | Total
-------------+---------------------+----------
Bihar | 519 187 | 706
| 73.51 26.49 | 100.00
Gujarat | 676 315 | 991
| 68.21 31.79 | 100.00
Kerala | 671 360 | 1,031
| 65.08 34.92 | 100.00
Odisha | 731 501 | 1,232
| 59.33 40.67 | 100.00
Punjab | 678 389 | 1,067
| 63.54 36.46 | 100.00
Rajasthan | 712 261 | 973
| 73.18 26.82 | 100.00
-------------+---------------------+----------
Total | 3,987 2,013 | 6,000
The same thing in pandas and in R
pandas
import pandas as pd pd.crosstab(df["state"], df["sector"])
R
library(dplyr) counts <- table(df$state, df$sector) counts # This R was run in a real R (4.6.1) and gave the same numbers.