Question

How does the mean absolute deviation (MAD) help you compare data sets?

Answer 1

Mean absolute deviation (MAD) is a simple measure of spread — it tells you, on average, how far the observations are from a chosen center (usually the mean or the median).

What it measures
- MAD = average of the absolute deviations from the center:
1. find the center (mean or median),
2. take absolute differences of each data point from that center,
3. average those absolute differences.
- The result is in the same units as the data and describes typical distance from the center.

How it helps compare data sets
- Compares variability: a smaller MAD means the data are more tightly clustered (more consistent); a larger MAD means more spread (more variability).
- Useful when two datasets have similar centers: MAD shows which one is more concentrated around its center.
- Works even if the datasets have different sizes (it averages over the number of observations).

Quick example
- Set A: 8, 9, 10, 11, 12 → mean = 10; deviations = 2,1,0,1,2 → MAD = (2+1+0+1+2)/5 = 1.2
- Set B: 5, 7, 10, 13, 15 → mean = 10; deviations = 5,3,0,3,5 → MAD = (5+3+0+3+5)/5 = 3.2
→ Set B is more spread out than Set A.

Caveats and tips
- If you want resistance to outliers, compute MAD about the median (median absolute deviation) rather than about the mean.
- If the datasets have very different means and you want to compare relative variability, use a normalized measure (e.g., coefficient of variation) instead.
- Use MAD together with measures of center (mean/median) and other spread measures (standard deviation, IQR) for a fuller picture.

Bottom line: MAD gives an easy-to-understand numeric summary of how spread out each dataset is, letting you directly compare their consistency or variability.