Demo Dataset

> df <- data.frame(ID = 1:20,
+                  NAME = LETTERS[1:20],
+                  STRATA = c(rep("S1",5), rep("S2", 8), rep("S3",7)))
> df
   ID NAME STRATA
1   1    A     S1
2   2    B     S1
3   3    C     S1
4   4    D     S1
5   5    E     S1
6   6    F     S2
7   7    G     S2
8   8    H     S2
9   9    I     S2
10 10    J     S2
11 11    K     S2
12 12    L     S2
13 13    M     S2
14 14    N     S3
15 15    O     S3
16 16    P     S3
17 17    Q     S3
18 18    R     S3
19 19    S     S3
20 20    T     S3

Stratified Sampling in Action

> library(dplyr) 
> set.seed(2024)  # setting seed for reproducibility
> df %>% 
+   group_by(STRATA) %>%  
+   sample_n(3)  # taking 3 samples from each stratum
# A tibble: 9 × 3
# Groups:   STRATA [3]
     ID NAME  STRATA
  <int> <chr> <chr> 
1     2 B     S1    
2     1 A     S1    
3     4 D     S1    
4     9 I     S2    
5    12 L     S2    
6     6 F     S2    
7    18 R     S3    
8    14 N     S3    
9    15 O     S3

A practical example

Sample size 128, strate 2 ( age <1 month and >1 month

> set.seed(2024)
> df_prac <- data.frame(ID = rnorm(128)*100 |> round(),
+                       AGE_GROUP = sample(c("Age < 1", "Age > 1"), size = 128, replace = TRUE))
> head(df_prac)
         ID AGE_GROUP
1  98.19694   Age < 1
2  46.87150   Age < 1
3 -10.79713   Age < 1
4 -21.28782   Age > 1
5 115.80985   Age > 1
6 129.23548   Age < 1
> df_prac %>% 
+   group_by(AGE_GROUP) %>%  # AGE_GROUP contains two strata
+   sample_n(10) # Taking 10 samples from each stratum
# A tibble: 20 × 2
# Groups:   AGE_GROUP [2]
        ID AGE_GROUP
     <dbl> <chr>    
 1 -124.   Age < 1  
 2  -76.6  Age < 1  
 3  111.   Age < 1  
 4   37.4  Age < 1  
 5  -49.9  Age < 1  
 6 -130.   Age < 1  
 7   68.5  Age < 1  
 8 -127.   Age < 1  
 9  -21.3  Age < 1  
10 -112.   Age < 1  
11   -2.89 Age > 1  
12  -18.3  Age > 1  
13  -32.9  Age > 1  
14 -120.   Age > 1  
15 -117.   Age > 1  
16 -327.   Age > 1  
17  -12.9  Age > 1  
18  197.   Age > 1  
19   49.2  Age > 1  
20  -26.5  Age > 1
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