1.5. SAS PROCEDURES -> Equivalent in R

1.5.3. PROC MEANS - Compute Descriptive Statistics in SAS and R

**Example Input Table:** `scores`

**Goal**

Compute descriptive statistics for SCORE overall and within each SUBJECT:

• Count (n)

• Mean, standard deviation

• 25th and 75th percentiles (q1, q3)

• Number of missing values

1. Overall Summary

SAS Code

proc means data=scores noprint;
  var score;
  output n=n mean=mean std=sd q1=q1 q3=q3 nmiss=nmissing out=overall_stats;
run;

Explanation:

• proc means: Starts the summary procedure.

• noprint: Suppresses printing the output to the Results window.

• var score;: Specifies the variable to summarize.

• output ... out=...: Stores summary results (n, mean, std, etc.) into a new dataset named overall_stats.

R Code

# Dummy data for execution
library(dplyr)
scores <- tibble::tibble(
  STUDENT = c("A001","A002","A003","A004","A005","A006","A007","A008","A009","A010","A011","A012","A013","A014","A015"),
  SUBJECT = c("Math","Math","Math","Math","Math","Science","Science","Science","Science","Science","History","History","History","History","History"),
  SCORE = c(78,85,91,62,74,83,77,65,90,NA,69,74,80,85,77)
)
overall_stats <- scores %>%
  dplyr::summarize(
    n = n(),
    mean = mean(SCORE, na.rm = TRUE),
    sd = sd(SCORE, na.rm = TRUE),
    q1 = quantile(SCORE, 0.25, na.rm = TRUE),
    q3 = quantile(SCORE, 0.75, na.rm = TRUE),
    nmissing = sum(is.na(SCORE))
  )

Explanation:

• summarize(): Reduces the entire data frame to a single-row summary.

• n(): Counts rows (non-missing).

• mean(), sd(): Basic statistics (exclude NAs).

• quantile(..., 0.25 / 0.75): Computes the 25th and 75th percentiles.

• sum(is.na()): Counts how many values are missing in the column.

Output Table

2. Summary by SUBJECT

SAS Code

proc sort data=scores; by subject;

proc means data=scores noprint;
  by subject;
  var score;
  output n=n mean=mean std=sd q1=q1 q3=q3 nmiss=nmissing out=by_subject_stats;
run;

Explanation:

• proc sort ... by subject;: Required step before using BY in SAS.

• BY SUBJECT: Tells SAS to compute statistics separately for each group.

• output ... out=by_subject_stats;: Saves the results for each SUBJECT to a new dataset.

R Code

by_subject_stats <- scores %>%
  group_by(SUBJECT) %>%
  dplyr::summarize(
    n = n(),
    mean = mean(SCORE, na.rm = TRUE),
    sd = sd(SCORE, na.rm = TRUE),
    q1 = quantile(SCORE, 0.25, na.rm = TRUE),
    q3 = quantile(SCORE, 0.75, na.rm = TRUE),
    nmissing = sum(is.na(SCORE))
  )

Explanation:

• group_by(SUBJECT): Groups the data frame by subject category.

• summarize(...): Applies each summary function to the grouped values.

Output Table

3. Alternatives to dplyr::summarize()

**What More Can You Explore**

1. Add More Summary Metrics

• min(SCORE), max(SCORE), IQR(), median(), var()
• Helps validate outliers or spread

2. Multi-Grouping

• Add group_by(SUBJECT, GENDER) to replicate BY across multiple class variables.

3. Visualize Results

• Boxplots, histograms, or violin plots for score distributions
• Plot mean ± sd error bars by subject

4. Export to Reports

• Use gt::gt(), flextable, or kable() for presentation-ready summary tables
• Save as Excel or HTML

5. Build a Custom summary_table() Function Reusable for any dataset:

summary_table <- function(data, var, group = NULL) {
  if (!is.null(group)) data <- data %>% group_by(across(all_of(group)))
  data %>%
    dplyr::summarize(
      n = sum(!is.na(.data[[var]])),
      mean = mean(.data[[var]], na.rm = TRUE),
      sd = sd(.data[[var]], na.rm = TRUE),
      q1 = quantile(.data[[var]], 0.25, na.rm = TRUE),
      q3 = quantile(.data[[var]], 0.75, na.rm = TRUE),
      nmiss = sum(is.na(.data[[var]])),
      .groups = "drop"
    )
}

overall_stats <- summary_table(scores, "SCORE")

**Resource download links**

1.5.3.-PROC-MEANS-Compute-Descriptive-Statistics-in-SAS-and-R.zip