What is the difference between the Outlier Rule and the Empirical Rule for detecting outliers?

The Empirical Rule (68–95–99.7) applies only to normally distributed data and flags any value beyond 3 standard deviations as an outlier. The Outlier Rule uses quartiles and the 1.5 \times IQR multiplier, which works on any shaped distribution and is more resistant to skew. Choose the Empirical Rule if your data is clearly bell-shaped; use the Outlier Rule for most real-world datasets with unknown distributions.

Why 1.5? Where does that specific number come from?

The constant 1.5 is an empirical choice popularized by John Tukey and assumes roughly normal data. It is designed to flag roughly the outer 0.7\% of values in a normal distribution while ignoring typical variation. Other multipliers (like 1 or 3) are sometimes used for different sensitivity; 1.5 is the standard default in statistics courses and software packages.

Can the lower boundary be negative if my data is all positive numbers?

Yes. If your dataset has low values and tight clustering, Q_1 - 1.5 \cdot IQR can easily be negative even though all actual data points are positive (like times or test scores). A negative boundary simply means no low outliers can exist—any value still below that boundary would have to be even more negative, which is impossible in your dataset.

If I find an outlier, should I always remove it from my data?

No. Finding an outlier is a signal to investigate, not a signal to delete. Some outliers are errors (typos or faulty sensors), some are real but unusual (a record-breaking performance), and some reveal important information about your process. Always check the source before deciding whether to exclude it from further analysis.

Outlier Rule

The Outlier Rule identifies data values farther than 1.5 times the interquartile range from the quartiles—use it to flag unusual observations.

x < Q_1 - 1.5\,\text{IQR} \quad \text{or} \quad x > Q_3 + 1.5\,\text{IQR}

Solve a problem with Outlier Rule

Type the problem. The solver will use Outlier Rule where Outlier Rule is the right tool, and tell you when it is not.

What each symbol means

What Outlier Rule takes

Symbol	Meaning
$x$	The specific data value being tested; $x$ must be in the same units as the rest of the dataset.
$Q_1$	The first quartile or 25th percentile, the median of the lower half of ordered data; if $Q_1$ equals $Q_3$ , the data lacks sufficient spread for reliable outlier detection.
$Q_3$	The third quartile or 75th percentile, the median of the upper half of ordered data; the distance from $Q_1$ to $Q_3$ (the IQR) determines how far a value must be to qualify as an outlier.

When to use it

When examining raw data to detect values that are unusually far from the central cluster and may warrant separate investigation.

Level

Usually taught in: Statistics

Worked examples

1. Identify a single outlier in a small dataset

Problem

Is 25 an outlier in the dataset 2, 5, 6, 7, 8, 9, 10, 25?

$Q_1 = \frac{5 + 6}{2} = 5.5$
Find the first quartile by averaging the 2nd and 3rd values of the ordered dataset.
$Q_3 = \frac{9 + 10}{2} = 9.5$
Find the third quartile by averaging the 6th and 7th values.
$IQR = 9.5 - 5.5 = 4$
Calculate the interquartile range, which is the spread of the middle half of the data.
$Q_3 + 1.5 \cdot IQR = 9.5 + 1.5(4) = 15.5$
The upper boundary is $Q_3 + 1.5 \times IQR = 15.5$ .
$x = 25 > 15.5$
Since 25 exceeds the upper boundary, it is an outlier by the Outlier Rule.

Answer: $x = 25 \text{ is an outlier}$

The dataset shows a clear separation between the main group (values 2–10) and the single large value (25). The Outlier Rule correctly flags 25 as unusual because it lies far beyond where we would expect typical variation in this dataset.

2. Detect a low outlier among mostly high scores

Problem

A student's test scores are 25, 78, 81, 83, 85, 87, 89, 92, 95. Is the score of 25 an outlier?

$Q_1 = \frac{78 + 81}{2} = 79.5$
Find the first quartile by taking the average of the 2nd and 3rd values in the ordered list.
$Q_3 = \frac{89 + 92}{2} = 90.5$
Find the third quartile by taking the average of the 7th and 8th values.
$IQR = 90.5 - 79.5 = 11$
Calculate the interquartile range, the spread of the middle 50% of scores.
$Q_1 - 1.5 \cdot IQR = 79.5 - 1.5(11) = 79.5 - 16.5 = 63$
The lower boundary is $Q_1 - 1.5 \times IQR = 63$ .
$x = 25 < 63$
Since 25 falls below the lower boundary, it qualifies as a low outlier.

Answer: $x = 25 \text{ is an outlier}$

The score of 25 is drastically lower than the student's remaining scores, which cluster between 78 and 95. The Outlier Rule detects this as an unusual value, which might indicate the student was ill, unprepared, or the test differed from the others in difficulty.

3. Identify unusual lap times during a training climb

Problem

A cyclist records nine lap times (in seconds) climbing a steep hill: 87, 92, 89, 95, 91, 88, 530, 90, 93. Which lap times are outliers?

$\text{Ordered data: } 87, 88, 89, 90, 91, 92, 93, 95, 530$
First, arrange the nine lap times in ascending order.
$Q_1 = 88$
With 9 data points, the first quartile is the value at position 2 after dividing into quarters.
$Q_3 = 93$
The third quartile is the value at position 8.
$IQR = 93 - 88 = 5$
The interquartile range represents typical variation in the cyclist's lap times.
$Q_3 + 1.5 \cdot IQR = 93 + 1.5(5) = 93 + 7.5 = 100.5$
Any time above 100.5 seconds exceeds the upper outlier threshold.
$x = 530 > 100.5$
The 530-second lap is far beyond the upper boundary, making it a clear outlier.

Answer: $x = 530 \text{ seconds is an outlier}$

One lap took 530 seconds (nearly 9 minutes) while all others took 87–95 seconds. This extreme outlier likely represents a mechanical failure, a rest break, or the cyclist walking part of the steep route. The Outlier Rule successfully identifies when a single performance measurement is radically different from the rest.

Common mistakes

Where Outlier Rule usually goes wrong

The mistake	Why it is wrong	The fix
Using $2 \times IQR$ instead of $1.5 \times IQR$ as the multiplier.	The $2 \times IQR$ rule (called Tukey's outer fence) is more conservative and flags fewer outliers; using it will miss genuine outliers that the standard $1.5 \times IQR$ rule would detect.	Always use exactly 1.5, not 2. The Outlier Rule specifically uses $Q_1 - 1.5 \cdot IQR$ and $Q_3 + 1.5 \cdot IQR$ as boundaries.
Forgetting to order the data before finding $Q_1$ and $Q_3$ .	Quartiles are defined by position in the sorted dataset; if the data is scrambled, $Q_1$ and $Q_3$ will be wrong, making the boundaries meaningless.	Always sort the data in ascending order first, then find $Q_1$ as the median of the lower half and $Q_3$ as the median of the upper half.
Swapping the boundaries by using $Q_1 - 1.5 \cdot IQR$ as the upper boundary and $Q_3 + 1.5 \cdot IQR$ as the lower boundary.	Since $Q_1$ is always less than $Q_3$ , reversing the boundaries places them on the wrong ends of your data; you will incorrectly mark normal values as outliers.	Remember: subtract $1.5 \cdot IQR$ from $Q_1$ for the lower boundary, and add $1.5 \cdot IQR$ to $Q_3$ for the upper boundary.

Tips and when to use something else

The multiplier 1.5 is a statistical convention chosen to work well with normally distributed data; roughly 0.7 $\%$ of values in a bell-shaped dataset will flag as outliers.
If $Q_1$ and $Q_3$ are equal (the middle 50 $\%$ of your data is identical), then $IQR = 0$ and both boundaries collapse to the same point—a sign your dataset lacks sufficient spread for reliable outlier detection.
For very small datasets (fewer than 5 values), quartiles are unstable and the Outlier Rule may not be reliable; use visual inspection or the $Z$ -score method instead.
Unlike methods that first remove extreme values, the Outlier Rule is robust because it uses quartiles—which are not affected by outliers themselves—to define the boundaries.

Related formulas

Interquartile RangeThe interquartile range measures the spread of the middle 50% of your data, telling you how compressed or spread out typical values are.Standard DeviationStandard deviation measures how spread out data points are from the average—use it to understand variability in a dataset.Z-ScoreThe Z-Score tells you how many standard deviations a data point is from the mean, revealing whether it is typical or unusual.MedianThe median is the middle value of ordered data, most useful when your data contains outliers that would skew the mean value.

Frequently asked questions

What is the difference between the Outlier Rule and the Empirical Rule for detecting outliers?: The Empirical Rule (68–95–99.7) applies only to normally distributed data and flags any value beyond 3 standard deviations as an outlier. The Outlier Rule uses quartiles and the $1.5 \times IQR$ multiplier, which works on any shaped distribution and is more resistant to skew. Choose the Empirical Rule if your data is clearly bell-shaped; use the Outlier Rule for most real-world datasets with unknown distributions.
Why 1.5? Where does that specific number come from?: The constant 1.5 is an empirical choice popularized by John Tukey and assumes roughly normal data. It is designed to flag roughly the outer 0.7 $\%$ of values in a normal distribution while ignoring typical variation. Other multipliers (like 1 or 3) are sometimes used for different sensitivity; 1.5 is the standard default in statistics courses and software packages.
Can the lower boundary be negative if my data is all positive numbers?: Yes. If your dataset has low values and tight clustering, $Q_1 - 1.5 \cdot IQR$ can easily be negative even though all actual data points are positive (like times or test scores). A negative boundary simply means no low outliers can exist—any value still below that boundary would have to be even more negative, which is impossible in your dataset.
If I find an outlier, should I always remove it from my data?: No. Finding an outlier is a signal to investigate, not a signal to delete. Some outliers are errors (typos or faulty sensors), some are real but unusual (a record-breaking performance), and some reveal important information about your process. Always check the source before deciding whether to exclude it from further analysis.

Need a different method?

The full solver is not scoped to one formula — type any problem and it will pick the method.

Open the math solver

Reviewed 2026-09-18