Standard deviation gets all the attention, but mean absolute deviation — MAD — is the spread measure you can explain in one sentence: on average, how far are the values from their mean? No squaring, no square roots, just distances. This mean absolute deviation calculator takes any data set and returns the MAD with the full working: the mean, each value’s distance from it, and the average of those distances.
How Mean Absolute Deviation Works
The procedure has three moves. Find the mean of the data. Measure each value's distance from that mean, dropping the sign — distance is distance, so −3 and +3 both count as 3; this is the "absolute" in the name, and it is what stops positive and negative deviations from cancelling each other out. Average the distances. That average is the MAD, expressed in the same units as the data itself: points, centimetres, dollars, milliseconds.
Worked Example: The Classic Data Set
Take 2, 4, 4, 4, 5, 5, 7, 9 — a set so well behaved it appears in textbooks. The mean is 40 ÷ 8 = 5. The absolute deviations are |2−5| = 3, then 1, 1, 1, 0, 0, 2, and |9−5| = 4, summing to 12. The MAD is 12 ÷ 8 = 1.5. Interpretation: a typical value in this set sits 1.5 units from the mean. (The standard deviation of the same set is exactly 2 — the next section explains why the two measures differ.)
MAD vs Standard Deviation
Both measure spread around the mean, but they weight distance differently. Standard deviation squares each deviation before averaging (then square-roots), which amplifies far-away values: in the worked example the outlier-ish 9 contributes 16 of the 32 squared units — half the total from one value. MAD weights every unit of distance equally, so it is smaller (1.5 vs 2.0) and less sensitive to extremes. Standard deviation dominates statistics because the squared form has elegant mathematical properties that make inference work. MAD dominates classrooms and quick reports because it is interpretable at a glance and honest about outliers. Neither is "correct"; they answer the same question with different sensitivities.
Why Some Fields Prefer MAD
Forecasters score predictions with mean absolute error — MAD's sibling — because a forecast off by twice as much is twice as bad, not four times as bad, and squared metrics let one terrible miss dominate the evaluation. Quality engineers like MAD-style measures for the same reason: they describe the typical miss, not the worst one magnified. In robust statistics, MAD's resistance to outliers makes it a building block for outlier detection itself (the modified z-score method uses a related median-based version). When a single wild value would wreck your conclusions, an absolute-deviation measure is the safer yardstick.
A Naming Trap: Two Different MADs
MAD also stands for median absolute deviation — the median of |x − median| — a different statistic used in robust outlier detection. This calculator computes the mean absolute deviation (average distance from the mean), which is what school curricula and most "MAD" homework mean. If your source is a robust-statistics text, check which center — mean or median — it averages distances from. The two values can differ substantially on skewed data.
Common Mistakes
Forgetting the absolute value is fatal: signed deviations from the mean always sum to zero, so a MAD computed without absolutes returns 0 for every data set — if your answer is 0, that is the bug. Mistake two is dividing by n − 1 out of standard-deviation habit; descriptive MAD divides by n, the count of values. Mistake three is reporting MAD for data whose center should be the median — pairing a mean-based spread with a median-based center mixes two stories about the same set.
A Quick Classroom Check
If you are computing MAD by hand for homework, there is a built-in error check: your deviations, before the absolute values are applied, must sum to zero, and after them, to n times the MAD. If either check fails, the mean is wrong or a distance was mismeasured. Teachers ask for MAD precisely because every intermediate step can be audited this way — there is nowhere for an arithmetic slip to hide.
Related Tools
Explore the Statistics Calculators hub for the companion measures, including the Median Calculator and the Standard Deviation Calculator.
Frequently Asked Questions
What is mean absolute deviation in simple words?
It is the average distance between each value in a data set and the set's mean. For 2, 4, 4, 4, 5, 5, 7, 9 the mean is 5 and the average distance from 5 is 1.5, so MAD = 1.5.
How is MAD different from standard deviation?
MAD averages the plain distances from the mean. Standard deviation squares the distances first, which punishes far-away values more — so standard deviation is almost always the larger number and reacts more strongly to outliers.
Do you divide by n or n − 1 for MAD?
By n. Mean absolute deviation is a descriptive summary of the data in hand, so it averages over all n values. The n − 1 adjustment belongs to sample variance and standard deviation.
Can MAD be zero?
Only when every value is identical — all distances from the mean are zero. Any variation at all gives a positive MAD. A MAD of exactly 0 on varied data means the absolute values were forgotten.
Is MAD the same as median absolute deviation?
No. This calculator averages distances from the mean. Median absolute deviation takes the median distance from the median — a robust-statistics tool with a different purpose. Check which one your source intends.
What is a good MAD value?
There is no universal scale — MAD inherits the data's units. Judge it relative to the mean and the context: a MAD of 1.5 points on a test averaging 80 is tight consistency; the same MAD on reaction times averaging 4 seconds would be chaos.
Why do the deviations from the mean sum to zero without absolute values?
By construction of the mean: it is the balance point of the data, so total distance above equals total distance below. Absolute values are what turn that balance into a measure of spread.