*51*

Developed by biostatistician Karl Pearson, **Pearson’s coefficient of skewness** is a way to measure the skewness in a sample dataset.

There are actually two methods that can be used to calculate Pearson’s coefficient of skewness:

**Method 1: Using the Mode**

Skewness = (Mean – Mode) / Sample standard deviation

**Method 2: Using the Median**

Skewness = 3(Mean – Median) / Sample standard deviation

In general, the second method is preferred because the mode is not always a good indication of where the “central” value of a dataset lies and there can be more than one mode in a given dataset.

The following step-by-step example shows how to calculate both versions of the Pearson’s coefficient of skewness for a given dataset in Excel.

**Step 1: Create the Dataset**

First, let’s create the following dataset in Excel:

**Step 2: Calculate the Pearson Coefficient of Skewness (Using the Mode)**

Next, we can use the following formula to calculate the Pearson Coefficient of Skewness using the mode:

The skewness turns out to be **1.295**.

**Step 3: Calculate the Pearson Coefficient of Skewness (Using the Median)**

We can also use the following formula to calculate the Pearson Coefficient of Skewness using the median:

The skewness turns out to be **0.569**.

**How to Interpret Skewness**

We interpret the Pearson coefficient of skewness in the following ways:

- A
**value of 0**indicates no skewness. If we created a histogram to visualize the distribution of values in a dataset, it would be perfectly symmetrical. - A
**positive value**indicates positive skew or “right” skew. A histogram would reveal a “tail” on the right side of the distribution. - A
**negative value**indicates a negative skew or “left” skew. A histogram would reveal a “tail” on the left side of the distribution.

In our previous example, the skewness was positive which indicates that the distribution of data values was positively skewed or “right” skewed.

**Additional Resources**

Check out this article for a nice explanation of left skewed vs. right skewed distributions.