Mean Absolute Deviation Calculator

What Does a Mean Absolute Deviation Calculator Do?

A mean absolute deviation calculator tells you how far, on average, the numbers in a data set sit from their center. Paste your values and it returns the MAD, the mean and median, the distance of every single value from the center, and a chart that shows where the data falls. If a set of quiz scores has a MAD of 5, a typical score landed about 5 points away from the average, on one side or the other. The variance calculator and the standard deviation calculator measure spread by squaring the distances instead of taking absolute values.

It handles the whole job in one place. You can measure the distances from the mean (the usual way), from the median, or from any number you type in. You can add a second data set and compare the two side by side, which is what most 6th and 7th grade problems ask for. And because “MAD” is also the short name for median absolute deviation, the calculator reports that figure as well, so you do not need a second tool.

The working is always on screen. Every row of the table shows the value, its distance from the center and the absolute distance, so you can check a homework answer line by line instead of trusting a single number.

What Is the Mean Absolute Deviation Formula?

MAD is the average of the absolute distances between each value and the center of the data:

Here xi is each value, x̄ is the mean, n is how many values there are, and the bars mean “take the absolute value,” which drops any minus sign. Σ tells you to add the distances up before dividing.

The absolute value is not decoration. Distances measured from the mean always cancel out: the values below the mean pull the total down by exactly as much as the values above it push it up, so the plain sum of deviations is always zero. Taking absolute values first stops that cancellation and leaves a number that reflects the real spread.

Two details are worth knowing. First, the answer comes out in the same units as your data, so a MAD of 8 minutes is 8 minutes, not 8 squared minutes. Second, the center does not have to be the mean. Some textbooks and software measure from the median instead, which gives the same formula with m (the median) in place of x̄. The calculator lets you pick either one, or a number of your own.

The separate statistic called the median absolute deviation uses a different recipe: find the median, take each value’s absolute distance from it, then take the median of those distances instead of the mean. It is covered in its own section below. The five number summary calculator is a close cousin, since it also builds on the median.

How Do You Use This Mean Absolute Deviation Calculator?

Type or paste your numbers and the results update as you go.

1. Enter data set A. Separate the numbers with commas, spaces, tabs or new lines. A column copied from Excel or Google Sheets pastes straight in. Negative numbers and decimals are fine, and up to 500 values are accepted.
2. Choose the center. Mean is the default and gives the standard mean absolute deviation. Median measures every distance from the middle value. Custom lets you type a target, such as a passing score of 70 or a forecast, and measures the average miss from that number.
3. Add a second data set if you want a comparison. Click “Compare with a second data set,” enter set B, and both appear in one table and on one chart with a shared scale.
4. Pick the decimal places from 2 to 8. The arithmetic always runs at full precision and only the display is rounded.

The quick buttons load ready-made examples: homework minutes, two bus routes, a data set with an outlier, and a short list to test by hand. Press Clear to go back to the starting example.

How Do You Calculate Mean Absolute Deviation by Hand?

Four steps: find the mean, subtract it from every value, drop the minus signs, and average what is left.

Take the five numbers 6, 8, 10, 14 and 17.

So the typical value in this set sits 3.6 away from the mean of 11. A quick check on step 2: the signed distances (−5, −3, −1, 3, 6) add up to zero, as they must. If yours do not, the mean is wrong or an arithmetic slip crept in.

If your class measures from the median instead, the median here is 10. The distances become 4, 2, 0, 4 and 7, which total 17, so the MAD around the median is 17 / 5 = 3.4. Same data, slightly different answer, because the center moved.

Example 1: How Do You Find the MAD of Homework Times?

Eight students report how many minutes they spent on homework last night: 22, 30, 35, 40, 41, 45, 52 and 55. The mean is 320 / 8 = 40 minutes.

On a typical night, a student’s homework time is about 8 minutes away from the 40-minute average. The population standard deviation for the same data is 10.27 minutes, a little larger than the MAD, which is the normal pattern (more on that below). The shaded band on the chart runs from 31.75 to 48.25 minutes, one MAD either side of the mean, and four of the eight students fall inside it.

Example 2: Two Data Sets With the Same Mean but Different Spread

This is the comparison that 6th and 7th grade standards ask for, and it is where MAD earns its keep. Two bus routes both average 30 minutes over eight mornings.

The averages are identical, so the mean alone cannot tell the routes apart. The MADs can. Route A is off by about one minute on a typical day. Route B is off by about seven and a half, which means that on a given morning you cannot predict B within a quarter of an hour. If you have to be at work by 9:00, A is the safer bus. The mean, median and mode calculator gives the averages in one step.

When the means do differ, a common way to compare is to express the gap in units of variability. If one class averages 78 with a MAD of 4 and another averages 70 with a MAD of 3, the means are 8 points apart, which is two times the larger MAD. Enter both sets in the calculator and it writes that comparison for you.

Example 3: What Happens to MAD When There Is an Outlier?

Take the values 12, 13, 14, 15, 16, 17 and 90. Six of them sit in a tidy cluster and one is far away. Here is how each measure reacts:

One extreme value multiplies the mean absolute deviation by roughly twelve, because the mean itself gets dragged toward the 90 and the 90 then contributes a huge distance. Measuring from the median softens the damage, and the median absolute deviation barely moves (1.5 to 2). The standard deviation is hit hardest of all, since it squares the large distance.

The practical rule: if your data might contain a stray extreme value, compare the mean-centered MAD with the median absolute deviation. A big gap between them is a sign that one or two points are doing most of the work. Look at the chart and decide whether those points are errors or real.

Mean Absolute Deviation vs Median Absolute Deviation: Which MAD Do You Need?

The letters MAD are used for two different statistics, and that confuses a lot of people. A calculator that reports one while your textbook means the other will give you the wrong answer with no warning.

If the question says “mean absolute deviation,” or comes from a middle-school or intro course, you want the left column, which is the headline result of this calculator. If it says “median absolute deviation,” or comes from a data-science or robust-statistics source, you want the right column, listed in the results as Median abs. deviation.

There is also a scaled version. Multiply the median absolute deviation by 1.4826 and it becomes a robust estimate of the standard deviation for roughly normal data. R’s mad() function applies that constant by default, which is why its answer is larger than the raw figure unless you set constant = 1. For the five numbers 6, 8, 10, 14 and 17 the raw median absolute deviation is 4 and the scaled one is 5.93.

Mean Absolute Deviation vs Standard Deviation: What Is the Difference?

Both describe spread, and both are in the same units as the data. The difference is in how they treat large distances. MAD adds up the plain distances. The standard deviation squares each distance first, then averages, then takes a square root, which gives big distances extra weight.

That has three consequences. The standard deviation is always at least as large as the MAD for the same data. A single outlier inflates the standard deviation more than the MAD. And the standard deviation has far nicer mathematical properties, which is why it sits underneath confidence intervals, z-scores and most of inferential statistics.

For data that follows a normal distribution the two are tied together by a fixed ratio: MAD is about 0.80 times the standard deviation, or equivalently the standard deviation is about 1.25 times the MAD. In Example 1 the ratio is 10.27 / 8.25 = 1.245, close to that benchmark. A ratio far above 1.25 hints at heavy tails or outliers. For a spread measure built from percentiles instead of distances, the quartile calculator reports the interquartile range. Exactly how the normal curve behaves is easy to explore with the normal curve calculator and the Gaussian distribution calculator, and the normal CDF calculator turns any mean and standard deviation into probabilities.

Which should you use? In school, use whichever your teacher or textbook names. In real analysis, the standard deviation is the default for modeling and testing, while MAD is a better plain-language summary (“the typical miss is X”) and holds up better when the data is messy.

How Do You Read and Interpret a MAD?

MAD answers one question: how far does a typical value sit from the center? Read it as a distance in the data’s own units.

• Zero means no spread. Every value equals the center. MAD can never be negative.
• Bigger means more spread. A MAD of 7.5 minutes describes less predictable data than a MAD of 1 minute, provided both are in minutes.
• Compare like with like. A MAD of 4 on exam scores out of 100 and a MAD of 4 on salaries in thousands say very different things. To compare across units or scales, divide the MAD by the mean. The calculator shows that as Relative MAD, so 8.25 on a mean of 40 is about 21%.
• The shaded band is a guide, not a promise. The chart shades one MAD either side of the center. For normal data that band holds about 57% of the values, but for skewed or clumpy data the share can be quite different.

There is no universal threshold for a “good” or “high” MAD. It only becomes meaningful next to something: another data set, a tolerance you care about, or the size of the mean.

Where Is Mean Absolute Deviation Used?

School statistics. The Common Core 6th grade standards name MAD as one of the measures of variability, and 7th grade asks students to compare two data sets using a measure of center plus a measure of variability. That is why the compare feature exists.

Forecasting and inventory. In operations management, MAD usually means the average absolute forecast error: how many units, on average, the forecast missed by, in either direction. It is easy to explain to a manager, and a common rule of thumb converts it to a standard deviation by multiplying by 1.25, which feeds safety-stock calculations. Enter your forecast errors as the data and pick a custom center of 0 to get exactly that number.

Quality control and measurement. The average distance of readings from a target or from the batch mean is a direct, readable gauge of consistency.

Finance and risk. Mean absolute deviation is sometimes used instead of variance as a measure of risk because it is less sensitive to extreme returns. (This is a description of a statistic, not investment advice.)

Comparing groups. Whenever two averages look the same, MAD tells you which group is steadier. If your data comes as values with counts or other weights, see the weighted mean calculator for the center first. Probability-based means appear throughout the binomial, Poisson and hypergeometric calculators.

How Do You Calculate MAD in Excel, Google Sheets, TI-84, R and Python?

Every tool follows the same recipe. Spreadsheets have a built-in for the mean-centered version, and the others take one short line.

Excel / Google Sheets   (data in A2:A9)
  =AVEDEV(A2:A9)                              mean absolute deviation (around the mean)
  =SUMPRODUCT(ABS(A2:A9-MEDIAN(A2:A9)))/COUNT(A2:A9)   around the median
  =MEDIAN(ABS(A2:A9-MEDIAN(A2:A9)))           median absolute deviation (array formula)

TI-84 Plus   (data in L1)
  mean(abs(L1-mean(L1)))

R
  x <- c(6, 8, 10, 14, 17)
  mean(abs(x - mean(x)))                      # 3.6
  mad(x, constant = 1)                        # 4  (median absolute deviation)

Python
  import numpy as np
  x = np.array([6, 8, 10, 14, 17])
  np.mean(np.abs(x - np.mean(x)))             # 3.6
  from scipy.stats import median_abs_deviation
  median_abs_deviation(x)                     # 4.0

On older Excel versions the median absolute deviation formula must be confirmed with Ctrl+Shift+Enter, while current versions accept it as a normal formula. A tip for pandas users: the old Series.mad() method was removed in pandas 2.0, so write the expression out instead. The calculator prints these lines again with your own numbers filled in when your list is short enough to display.

What Mistakes Do People Make When Finding MAD?

The first and most common is forgetting the absolute value. The signed distances add to zero, and an answer of exactly 0 for data that clearly varies is the giveaway.

Second, mixing up the two MADs. If your source defines the statistic as the median of absolute distances and you computed the average, the numbers will differ, sometimes by a lot (see Example 3).

Third, rounding the mean before subtracting. With a mean like 11.2857, rounding to 11.3 first shifts every distance a little, and the error survives into the final answer. Carry the full mean until the end.

Fourth, dividing by the wrong count. MAD divides by n, the number of values, even when your data is a sample. Some people reach for n − 1 out of habit from the sample standard deviation, but that adjustment is not part of the usual MAD definition.

Fifth, comparing MADs from data in different units, or reading a MAD of 3 as “small” without knowing the mean it belongs to.

Key Points to Remember

• MAD = Σ|x − center| / n. Find the center, take absolute distances, then average them.
• It is measured in the same units as your data and can never be negative.
• MAD is 0 only when every value is identical.
• “Mean absolute deviation” and “median absolute deviation” are different statistics. Check which one your course or software wants.
• One outlier can inflate the mean-centered MAD sharply. The median absolute deviation resists it.
• For normal data, standard deviation ≈ 1.25 × MAD.
• In Excel and Google Sheets, AVEDEV returns the mean absolute deviation.
• To compare spread across data sets, use the same units, or divide MAD by the mean.

Frequently Asked Questions

Further reading: NumPy documentation: numpy.mean, SciPy documentation: median_abs_deviation and NIST/SEMATECH e-Handbook of Statistical Methods.

Related Calculators

Need a different statistics tool? Try these, or browse the full list of free statistics calculators:

Weighted Mean CalculatorFind the center first when some values count more than others. Normal Curve CalculatorZ-scores, percentiles and normal / bell-curve probabilities. Normal CDF CalculatorArea under the normal curve between any two bounds. All StatsSolve CalculatorsBrowse the full directory of free statistics tools.