Central Limit Theorem: Formula, Interpretation, R, Python, SPSS and Excel Guide

What Is the Central Limit Theorem?

The Central Limit Theorem is one of the most important ideas in statistics. It explains why the sample mean is so useful in real data analysis. When repeated random samples are taken from a population, the means of those samples form their own distribution. This new distribution is called the sampling distribution of the mean.

The Central Limit Theorem says that as the sample size becomes larger, the sampling distribution of the mean becomes approximately normal, even when the original population is not perfectly normal. This is why many statistical methods can use normal-based reasoning for sample means, confidence intervals and hypothesis tests.

In this worked example, G3 final grade is used as the main population variable. The original G3 distribution is not perfectly normal because it includes low-grade values and a visible left-tail pattern. The absences variable is even more extreme because it is strongly right-skewed. However, when repeated sample means are generated, the sampling distributions become smoother and more normal-looking as sample size increases.

For related assumption-checking and normality topics, see the Q-Q Plot Normality Check, Shapiro-Wilk Test, Kolmogorov-Smirnov Test, DAgostino Pearson Test and One-Sample Z Test.

Central Limit Theorem Formula

The main Central Limit Theorem formula used in this guide is the standard error of the sample mean:

SE of sample mean = σ / √n

Where:

The formula explains the visual pattern in the charts. When n = 5, sample means vary more widely. When n = 100, sample means are much closer to the population mean. The standard error decreases because the population standard deviation is divided by the square root of the sample size.

Dataset and Clean SPSS-Ready Files Used

This Central Limit Theorem example uses the student-por.csv student performance dataset. The main variable is G3 final grade, and the second demonstration variable is absences. G3 shows how sample means stabilize for a grade variable, while absences shows that the Central Limit Theorem can still work even when the original variable is highly skewed.

External dataset source: UCI Machine Learning Repository: Student Performance dataset.

Verified SPSS, R and Python Results

The Central Limit Theorem workflow was reproduced in Python, R and SPSS. Python generated the clean SPSS-ready files and charts. R produced independent validation charts. SPSS imported the clean files and produced descriptive statistics, frequency tables, normality checks and sampling-distribution summaries.

Original Data Check

Original Distribution Shape

Sampling Distribution of G3 Sample Means

The main Central Limit Theorem result is visible in the G3 sample-mean table. The mean of repeated sample means stays close to the population mean, while the standard deviation of sample means becomes smaller as sample size increases.

Sampling Distribution of Absences Sample Means

The absences variable is strongly right-skewed in the raw data, but the sample means still become more stable as sample size increases. This is a strong practical demonstration of the Central Limit Theorem.

SPSS Output Transcript

Python Charts and Interpretation

1. Raw G3 Population Distribution

This chart shows the original G3 final-grade population distribution. The mean is about 11.91 and the median is 12.00. The distribution is not perfectly normal, but it is concentrated around the middle grade range. This chart is important because the Central Limit Theorem begins with the original population distribution before repeated sampling is performed.

2. Sampling Distribution of the Mean for G3

This is the core Central Limit Theorem chart. At n = 5, the sample means are spread more widely. At n = 30, the distribution is already much more stable and bell-shaped. At n = 100, sample means cluster tightly around the population mean. This visual pattern matches the standard error formula.

3. Standard Error Decreases as Sample Size Increases

The line chart shows that the simulated standard deviation of sample means closely follows the theoretical standard error σ / √n. This means the simulation is working correctly. As sample size increases, the standard error becomes smaller, so the sample mean becomes a more precise estimate of the population mean.

4. Raw Absences Distribution Is Right-Skewed

The raw absences variable is strongly right-skewed. Many students have zero or very few absences, while a smaller number have large absence values. The mean is higher than the median because the right-tail values pull the average upward. This chart is useful because it shows that the Central Limit Theorem is not limited to perfectly normal original data.

5. Central Limit Theorem with Skewed Absences Data

This chart demonstrates the practical power of the Central Limit Theorem. Even though the original absences distribution is highly skewed, the sampling distribution of the mean becomes more bell-shaped as sample size increases. At n = 5, the distribution still has visible skewness. At n = 50 and n = 100, the sample means are much more stable and closer to normal.

6. Q-Q Plots of G3 Sample Means

The Q-Q plots confirm what the histograms show. At smaller sample sizes, the tails depart more from the reference line. As sample size increases, the plotted points move closer to the line, meaning the sampling distribution becomes more normal. This is direct visual evidence for the Central Limit Theorem.

7. 95% Confidence Interval Coverage for G3

This chart connects the Central Limit Theorem with confidence intervals. The target coverage is 0.95. Smaller sample sizes show lower coverage, while larger samples move closer to the target. This is why normal-based confidence intervals usually perform better when sample size is larger and the sampling distribution is more stable.

8. Sample Mean Stabilizes as Sample Size Grows

This line chart shows convergence. Early sample means can jump up and down because small samples are unstable. As sample size grows, the sample mean fluctuates less and stays closer to the population mean. This is the same idea behind using larger samples for more reliable statistical estimates.

9. Population Distribution vs Sampling Distribution

This chart compares the raw population distribution with the sampling distribution. The population distribution is wider because it contains individual grades. The sampling distribution of means is narrower because each value is an average of many observations. This distinction is essential for understanding standard error.

R Validation Charts for the Central Limit Theorem

The R workflow produced matching validation charts. These charts confirm that the same Central Limit Theorem pattern appears outside Python and SPSS.

How to Run the Central Limit Theorem in Python, R, SPSS and Excel

Central Limit Theorem in Python

The Python workflow uses the cleaned dataset, saves charts, creates the clean SPSS import file and creates the sampling-means file used by SPSS.

import os
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from scipy import stats

base_dir = r"D:\DATA ANALYSIS\A Basic Descriptive Statistics Guides\Central Limit Theorem"
input_clean_csv = os.path.join(base_dir, "spss_ready_data.csv")

python_dir = os.path.join(base_dir, "Python")
spss_dir = os.path.join(base_dir, "SPSS")
os.makedirs(python_dir, exist_ok=True)
os.makedirs(spss_dir, exist_ok=True)

df = pd.read_csv(input_clean_csv)

required = ["school", "sex", "age", "studytime", "failures", "absences", "G1", "G2", "G3"]
missing = [c for c in required if c not in df.columns]
if missing:
    raise ValueError(f"Missing required columns: {missing}")

clean = df[required].copy()
for col in ["age", "studytime", "failures", "absences", "G1", "G2", "G3"]:
    clean[col] = pd.to_numeric(clean[col], errors="coerce")

clean = clean.dropna(subset=["G3", "absences"]).reset_index(drop=True)
clean.insert(0, "case_id", np.arange(1, len(clean) + 1))

clean.to_csv(os.path.join(spss_dir, "clt_clean_data_for_spss.csv"), index=False)

rng = np.random.default_rng(12345)
sample_sizes = [5, 10, 30, 50, 100]
n_sim = 5000

wide = pd.DataFrame({"simulation_id": np.arange(1, n_sim + 1)})

for n in sample_sizes:
    g3_means = [rng.choice(clean["G3"], size=n, replace=True).mean() for _ in range(n_sim)]
    abs_means = [rng.choice(clean["absences"], size=n, replace=True).mean() for _ in range(n_sim)]
    wide[f"g3_mean_n{n}"] = g3_means
    wide[f"absences_mean_n{n}"] = abs_means

wide.to_csv(os.path.join(spss_dir, "clt_sampling_means_wide_for_spss.csv"), index=False)

print("Population n:", len(clean))
print("G3 population mean:", clean["G3"].mean())
print("G3 population SD:", clean["G3"].std(ddof=1))
print("Absences population mean:", clean["absences"].mean())
print("Absences population SD:", clean["absences"].std(ddof=1))
print("SPSS files saved in:", spss_dir)

Central Limit Theorem in R

The R workflow reads the same cleaned data, simulates sample means and saves the R output charts inside the topic folder.

install.packages(c("tidyverse"))

library(tidyverse)

base_dir <- "D:/DATA ANALYSIS/A Basic Descriptive Statistics Guides/Central Limit Theorem"
input_clean_csv <- file.path(base_dir, "spss_ready_data.csv")
r_dir <- file.path(base_dir, "R")
dir.create(r_dir, showWarnings = FALSE, recursive = TRUE)

set.seed(12345)

student <- read.csv(input_clean_csv, stringsAsFactors = FALSE)

clean <- student %>%
  mutate(
    case_id = row_number(),
    G3 = as.numeric(G3),
    absences = as.numeric(absences),
    G1 = as.numeric(G1),
    G2 = as.numeric(G2),
    age = as.numeric(age),
    studytime = as.numeric(studytime),
    failures = as.numeric(failures)
  ) %>%
  filter(!is.na(G3), !is.na(absences))

sample_sizes <- c(5, 10, 30, 50, 100)
n_sim <- 5000

simulate_means <- function(x, n, n_sim = 5000) {
  replicate(n_sim, mean(sample(x, size = n, replace = TRUE)))
}

summary_table <- tibble()

for (n in sample_sizes) {
  g3_means <- simulate_means(clean$G3, n, n_sim)
  abs_means <- simulate_means(clean$absences, n, n_sim)

  summary_table <- bind_rows(
    summary_table,
    tibble(variable = "G3", sample_size = n, mean = mean(g3_means), sd = sd(g3_means)),
    tibble(variable = "absences", sample_size = n, mean = mean(abs_means), sd = sd(abs_means))
  )
}

print(summary_table)

write.csv(summary_table, file.path(r_dir, "clt_r_summary_table.csv"), row.names = FALSE)

Central Limit Theorem in SPSS

The SPSS syntax below imports the two Python-generated clean files. The first file checks the original data. The second file checks the sampling distributions.

* ============================================================.
* Central Limit Theorem - SPSS Syntax.
* Run Python first so SPSS can use:
* 1) clt_clean_data_for_spss.csv
* 2) clt_sampling_means_wide_for_spss.csv
* ============================================================.

SET UNICODE=ON.
SET DECIMAL=DOT.
SET PRINTBACK=ON.
SET TNUMBERS=VALUES.
SET TVARS=LABELS.

GET DATA
 /TYPE=TXT
 /FILE="D:\DATA ANALYSIS\A Basic Descriptive Statistics Guides\Central Limit Theorem\SPSS\clt_clean_data_for_spss.csv"
 /ENCODING='UTF8'
 /DELCASE=LINE
 /DELIMITERS=","
 /QUALIFIER='"'
 /ARRANGEMENT=DELIMITED
 /FIRSTCASE=2
 /IMPORTCASE=ALL
 /VARIABLES=
 case_id F8.0
 school A20
 sex A20
 age F8.2
 studytime F8.2
 failures F8.2
 absences F8.2
 G1 F8.2
 G2 F8.2
 G3 F8.2.
CACHE.
EXECUTE.

DATASET NAME CLTOriginalData WINDOW=FRONT.

TITLE "Central Limit Theorem: Original Data Check".

DESCRIPTIVES VARIABLES=G1 G2 G3 absences age studytime failures
 /STATISTICS=MEAN STDDEV MIN MAX.

FREQUENCIES VARIABLES=G3 absences
 /STATISTICS=MEAN MEDIAN STDDEV VARIANCE MINIMUM MAXIMUM SKEWNESS SESKEW KURTOSIS SEKURT
 /PERCENTILES=25 50 75
 /ORDER=ANALYSIS.

EXAMINE VARIABLES=G3 absences
 /PLOT BOXPLOT HISTOGRAM NPPLOT
 /COMPARE GROUPS
 /STATISTICS DESCRIPTIVES EXTREME
 /CINTERVAL 95
 /MISSING LISTWISE
 /NOTOTAL.

GET DATA
 /TYPE=TXT
 /FILE="D:\DATA ANALYSIS\A Basic Descriptive Statistics Guides\Central Limit Theorem\SPSS\clt_sampling_means_wide_for_spss.csv"
 /ENCODING='UTF8'
 /DELCASE=LINE
 /DELIMITERS=","
 /QUALIFIER='"'
 /ARRANGEMENT=DELIMITED
 /FIRSTCASE=2
 /IMPORTCASE=ALL
 /VARIABLES=
 simulation_id F8.0
 g3_mean_n5 F12.6
 g3_mean_n10 F12.6
 g3_mean_n30 F12.6
 g3_mean_n50 F12.6
 g3_mean_n100 F12.6
 absences_mean_n5 F12.6
 absences_mean_n10 F12.6
 absences_mean_n30 F12.6
 absences_mean_n50 F12.6
 absences_mean_n100 F12.6.
CACHE.
EXECUTE.

DATASET NAME CLTSamplingMeans WINDOW=FRONT.

TITLE "Central Limit Theorem: Sampling Distributions of the Mean".

DESCRIPTIVES VARIABLES=
 g3_mean_n5 g3_mean_n10 g3_mean_n30 g3_mean_n50 g3_mean_n100
 absences_mean_n5 absences_mean_n10 absences_mean_n30 absences_mean_n50 absences_mean_n100
 /STATISTICS=MEAN STDDEV MIN MAX.

FREQUENCIES VARIABLES=
 g3_mean_n5 g3_mean_n10 g3_mean_n30 g3_mean_n50 g3_mean_n100
 absences_mean_n5 absences_mean_n10 absences_mean_n30 absences_mean_n50 absences_mean_n100
 /STATISTICS=MEAN MEDIAN STDDEV VARIANCE MINIMUM MAXIMUM SKEWNESS SESKEW KURTOSIS SEKURT
 /PERCENTILES=25 50 75
 /ORDER=ANALYSIS.

OUTPUT EXPORT
 /CONTENTS EXPORT=VISIBLE
 /PDF DOCUMENTFILE="D:\DATA ANALYSIS\A Basic Descriptive Statistics Guides\Central Limit Theorem\SPSS\Central-Limit-Theorem-SPSS-output.pdf".

Central Limit Theorem in Excel

Excel can demonstrate the main Central Limit Theorem formula and standard error calculation. It is useful for teaching the idea, but Python and R are better for repeated simulations.

Central Limit Theorem standard error in Excel:
=STDEV.S(B2:B650)/SQRT(30)

95% confidence interval lower bound:
=AVERAGE(B2:B31)-1.96*(STDEV.S(B2:B31)/SQRT(30))

95% confidence interval upper bound:
=AVERAGE(B2:B31)+1.96*(STDEV.S(B2:B31)/SQRT(30))

How to Report the Central Limit Theorem Result

A strong report should mention the original population variable, sample sizes, number of simulations, population mean, population standard deviation, standard error formula and the observed decrease in the spread of sample means.

Related Data Analysis Guides

The Central Limit Theorem supports many other statistical methods. After reading this guide, continue with One-Sample Z Test, One-Proportion Z Test, Q-Q Plot Normality Check, Shapiro-Wilk Test, Kolmogorov-Smirnov Test and Box Plot Interpretation. If you are checking model assumptions after using means in regression, also review Goldfeld-Quandt Test, Brown-Forsythe Test and Influence Diagnostics.

Common Mistakes

1. Thinking the Central Limit Theorem makes raw data normal

The Central Limit Theorem does not make the original raw variable normal. In this example, raw G3 and raw absences both failed normality tests. The theorem applies to the sampling distribution of the mean.

2. Confusing standard deviation with standard error

Standard deviation describes the spread of individual observations. Standard error describes the spread of sample means. As sample size increases, standard error decreases.

3. Ignoring sample size

Sample size is central to the Central Limit Theorem. Very small samples may still produce sampling distributions with visible skewness or tail problems, especially when the original population is strongly skewed.

4. Running SPSS before creating clean files

For this workflow, Python should create clt_clean_data_for_spss.csv and clt_sampling_means_wide_for_spss.csv first. SPSS should then import those clean files.

5. Reporting p = .000

SPSS may display very small p-values as .000. In final statistical writing, report these as p < .001, not p = .000.

Download SPSS Output and Verification Files

The SPSS output PDF verifies the original data check, descriptive statistics, normality tests, frequency tables, sampling-distribution summaries and Central Limit Theorem interpretation.

External References for the Central Limit Theorem

This post uses verified Python, R and SPSS outputs together with external statistical references and software documentation.

• UCI Machine Learning Repository: Student Performance dataset
• IBM SPSS Statistics documentation
• The Comprehensive R Archive Network
• SciPy statistical functions documentation
• Pandas documentation