LESSON: Impact of Outliers on Measures of Center

Identifying Outliers

What is an outlier? An outlier is a data point that is much higher or much lower than the main body of our data set. 

When we order our data from least to greatest it becomes apparent that some of our data points might seem much higher or lower than the rest of the data points. Visually inspecting the data in this way is the first step to identifying outliers.

Visual inspection is subjective though. I might determine a data point is an outlier, while you may argue that it is not. 

Statisticians have come up with a general rule for determining if points are outliers. This rule is not exact science, but it's the rule used across the board. To use this rule you first have to know how to calculate the Interquartile Range (or IQR). The IQR is the difference between quartile 1 and quartile 3. 

Watch this video to see how the IQR is calculated and how to use the rule to find outliers:

Practice Identifying Outliers

Practice using the IQR rule for identifying outliers by following the steps on this webpage: 

Khan Academy - Identifying Outliers with the 1.5xIQR Rule

Removing an Outlier

Outliers can have an interesting impact on our calculations for the measure of center. Oftentimes, outliers are "flukes", "mistakes", or come about by pure luck. We don't want those types of situations to have an impact on our measures of center, so we remove them!

In this video, it is determined by visual inspection that there is an outlier in the data set. Watch what happens to the calculations for the measure of center when the outlier is removed.

Increasing an Outlier

In this lesson video, you will see what happens to the calculations for the measures of center when an outlier is increased. When the video is over, see if you can determine what would happen to the measures of center if instead, a low outlier was decreased.