Winsorized Mean Calculator

Calculator Input

Example Data Table

Worked Example

This example uses 12.5% lower and 12.5% upper winsorization. One lowest and one highest value are capped.

Position	Original Sorted Value	Winsorized Value	Status
1	12	13	Lower cap
2	13	13	Unchanged
3	14	14	Unchanged
4	15	15	Unchanged
5	16	16	Unchanged
6	18	18	Unchanged
7	19	19	Upper cap reference
8	100	19	Upper cap

Raw mean: 25.875

Winsorized mean: 15.875

Interpretation: The extreme outlier no longer dominates the average, so the central tendency becomes more stable and representative.

Formula Used

The winsorized mean replaces extreme tail observations with cutoff values based on the chosen lower and upper percentages.

Step 1: Sort the dataset from smallest to largest.

Step 2: Compute replacement counts.

k_L = floor(n × p_L / 100)

k_U = floor(n × p_U / 100)

Here, n is the number of observations, p_L is the lower percentage, and p_U is the upper percentage.

Step 3: Replace the smallest k_L values with the next retained lower value.

Step 4: Replace the largest k_U values with the next retained upper value.

Winsorized Mean = (Σ winsorized values) / n

This calculator also reports the raw mean, trimmed mean, median, variance, standard deviation, and the exact change caused by winsorization.

How to Use This Calculator

Enter a label for the dataset.
Paste numeric values into the data box.
Set the lower and upper winsorization percentages.
Choose the number of displayed decimal places.
Click Calculate Winsorized Mean.
Review the result cards and observation detail table.
Download the summary as CSV or PDF when needed.

Frequently Asked Questions

1) What is a winsorized mean?

A winsorized mean replaces extreme observations with nearby cutoff values before averaging. It reduces outlier influence while still keeping every record in the dataset.

2) How is it different from a trimmed mean?

A trimmed mean removes tail observations entirely. A winsorized mean keeps the same sample size, but caps extreme values at selected boundary positions.

3) When should I use winsorization?

Use it when a few extreme values distort the average, but you still want every observation represented in the final calculation and summary statistics.

4) Why do percentages sometimes adjust zero observations?

The calculator uses floor counts. In small samples, low percentages may produce zero capped values because the tail count rounds down.

5) Can lower and upper percentages be different?

Yes. Asymmetric winsorization is useful when one tail contains stronger outliers than the other or when domain rules justify uneven treatment.

6) Does winsorization change variance and standard deviation?

Yes. Capping extreme values usually reduces spread, so the winsorized variance and standard deviation are often smaller than the raw statistics.

7) Can I use negative values or decimals?

Yes. The calculator accepts integers, decimals, and negative numbers, provided every entry is numeric and separated clearly.

8) What does the observation detail table show?

It lists each sorted value, its winsorized replacement, the numeric adjustment, and whether the point stayed unchanged or was capped.