StatReason

Mean Absolute Deviation Calculator

Find the average distance each value sits from a center point, ignoring direction.

What this answers

Mean absolute deviation answers "on average, how far does each value sit from the center?" using original units directly, without the squaring that standard deviation applies. This makes it more directly interpretable in plain terms, though standard deviation remains far more common in formal statistical work because of its mathematical properties.

How it is calculated

Each value's absolute distance from a chosen center (the mean, by default) is computed, and those distances are averaged. Using the mean as center is standard, but you can supply any center you choose, such as a target value or a benchmark, and the calculator will compute distances from that instead.

Worked example

For the values 1, 2, and 3 with the default center (the mean, which is 2): the absolute distances are 1, 0, and 1, averaging to about .667. Compare this to the sample standard deviation of the same data, which is exactly 1, illustrating how these two spread measures can give noticeably different numbers even on identical data, since squaring in the standard deviation formula weights larger distances more heavily than the simple averaging used here.

Assumption audit

Calculated from your data: the center used (the mean, unless you supplied your own) and the resulting average absolute distance.
Evidence to review: whether the mean is actually a meaningful center for your data, or whether a different center (a target value, or the median) would better represent the comparison you intend.
You must verify: that averaging absolute distances is the right way to summarize spread for your specific question, since this measure treats every deviation equally regardless of size, unlike variance-based measures.

Common mistakes

Assuming mean absolute deviation and standard deviation will be roughly equal is a common misconception; for normally distributed data they differ by a known, fixed ratio, but for skewed or heavy-tailed data that relationship breaks down, and the two measures can tell noticeably different stories about the same dataset's spread.

Limitations

Mean absolute deviation is used far less often than standard deviation in formal statistical modeling, since it lacks some of the mathematical properties (like differentiability) that make variance-based measures more tractable for further analysis.