I try to do a "simple" exact 1 to 1 matching. I would expect that the resulting data consist of two groups with the same size. But in my MWE it looks like this where the groups are not of thes same size.
Sample Sizes:
Control Treated
All 48. 52
Matched (ESS) 16.66 31
Matched 34. 31
Unmatched 14. 21
Discarded 0. 0
I'm sure I misunderstand something here.
Here is an MWE constructing sample data.
# see: https://cran.r-project.org/web/packages/MatchIt/vignettes/MatchIt.html
library("MatchIt")
set.seed(0)
df <- data.frame(
group = sample(c('X', 'Y'), 100, replace = TRUE),
foo = sample(LETTERS[1:4], 100, replace = TRUE),
bar = sample(10:20, 100, replace = TRUE)
)
m.out <- matchit(
group ~ foo + bar,
data = df,
method = "exact",
)
summary(m.out)
# get data
m.data <- match.data(m.out)
head(m.data)
The head of the resulting data.
> head(m.data)
group foo bar weights subclass
1 Y A 18 1.0000000 1
2 X D 19 1.4623656 2
3 Y D 19 1.0000000 2
5 X A 11 0.5483871 4
8 X C 11 1.0967742 17
14 X C 17 0.5483871 16
In my expectation the resulting data should look like this, where each pair has the same value in foo and bar.
group foo bar
1 Y A 18
37 X A 18
2 X D 19
3 Y D 19
5 X A 11
28 Y A 11
6 Y A 12
10 Y A 12
7 X A 10
92 X A 10
...
