Table of Contents
What Is the Lower Fence?
The lower fence is a boundary used to identify potential outliers in the lower end of a dataset. Any data point below the lower fence is considered a statistical outlier. It is a key component of the box-and-whisker plot methodology developed by John Tukey.
The lower fence is calculated using the first quartile (Q1) and the interquartile range (IQR). The standard lower fence uses a multiplier of 1.5 times the IQR, while the extreme lower fence uses 3 times the IQR to identify only extreme outliers.
Formula
Where k = 1.5 for standard outlier detection and k = 3.0 for extreme outlier detection.
Examples
| Dataset | Q1 | Q3 | IQR | Lower Fence |
|---|---|---|---|---|
| 1,3,5,7,9,11,13 | 3 | 11 | 8 | -9 |
| 10,20,30,40,50 | 15 | 45 | 30 | -30 |
Frequently Asked Questions
What does a negative lower fence mean?
A negative lower fence simply means the outlier threshold is below zero. If all your data values are positive, a negative lower fence means there are no low-end outliers. The fence itself is just a threshold, not a data value.
Should I always remove outliers?
Not necessarily. Outliers may represent genuine data points (e.g., exceptional performance), data entry errors, or measurement anomalies. Investigate each outlier before deciding to remove or retain it.
Why use 1.5 as the multiplier?
John Tukey chose 1.5 because for normally distributed data, approximately 0.7% of values fall outside 1.5 times the IQR. This provides a good balance between identifying true outliers and avoiding false positives.