find the mean absolute deviation and use it to explain how two data sets with the same mean can be very different.
Phase 4: Build
Do Now. Back to our unit puzzle: two classes have the same mean test score. Can they still be very different? Make a prediction.
Two small sample classes each have a mean quiz score of 16 out of 20. Must the two classes have similar scores?
The reading opens over the lesson; close it to come back to exactly where you were.
Spread around the mean. The mean absolute deviation (MAD) is the average distance between each value and the mean. A small MAD means the values huddle close to the mean. A large MAD means they are spread far apart.
Four steps. (1) Find the mean. (2) Find each value’s distance from the mean. A distance is never negative; it is an absolute value. (3) Add the distances. (4) Divide by the number of values.
Worked example 1: Ava’s Class A. Eight sample quiz scores out of 20: 14, 15, 15, 16, 16, 17, 17, 18. Sum 128, so the mean is 128 ÷ 8 = 16. Distances from 16: 2, 1, 1, 0, 0, 1, 1, 2. Sum of distances: 8. MAD = 8 ÷ 8 = 1 point. On average, a score is 1 point from the mean.
Worked example 2: Ben’s Class B. Eight sample scores: 10, 11, 13, 16, 18, 20, 20, 20. Sum 128, so the mean is also 16. Distances from 16: 6, 5, 3, 0, 2, 4, 4, 4. Sum: 28. MAD = 28 ÷ 8 = 3.5 points. On average, a score is 3.5 points from the mean.
Answering the puzzle. Both classes have a mean of 16, but Class B’s MAD is 3.5 times Class A’s. Class A’s scores are all close together; Class B has some students scoring 10 or 11 and three with perfect scores. A teacher would plan very differently for these two classes. The mean alone hides that.
Which spread goes with which center? The MAD measures spread around the mean, so report them together. The IQR measures the spread of the middle half around the median, so report it with the median. Class B’s median is 17 and its IQR is 20 − 12 = 8; Class A’s median is 16 and its IQR is 17 − 15 = 2.
Use a table with three columns: value, distance from the mean, running total.
Measure the spread for your proposal.
Sample data: two garden boxes were checked for ripe tomatoes on 5 days. Box 1: 6, 7, 7, 7, 8. Box 2: 3, 5, 7, 9, 11. Find the mean and the MAD of each. Which box gave a more consistent harvest? Explain.
Find the MAD of your team's survey data. Write one sentence that reports the mean and the MAD together, in context.
The mean tells where; the MAD tells how spread out.
Exit ticket: Sam says, “The MAD of 10, 10, 10, 30 is 20, because 30 − 10 = 20.” What did he find instead? What is the MAD?
Pacing: warm-up and reading 7 minutes, mini lecture 11, guided practice 14, independent 6, closure 2.
Likely misconceptions: keeping negative signs on deviations (so they add to 0); dividing by the number of nonzero distances; confusing MAD with range. Terminology alert: the NIST page uses the letters MAD for the median absolute deviation, a different measure. In grade 6 and on the PSSA, MAD means mean absolute deviation (NIST’s “average absolute deviation”). Point this out if students open the page.
Supports: a number line with the mean marked so students count distances; calculators for the final division. Stretch: create a data set of 4 values with mean 10 and MAD 0.5 (for example 9, 10, 10, 11).
Answer guide: MAD 2 (distances 3, 1, 1, 3); 0; Class A is closer; false; the range; 2.4. Independent 1: Box 1 mean 35 ÷ 5 = 7, distances 1, 0, 0, 0, 1, MAD 2 ÷ 5 = 0.4; Box 2 mean 7, MAD 2.4; Box 1 is more consistent. Exit ticket: Sam found the range; the mean is 60 ÷ 4 = 15, distances 5, 5, 5, 15, MAD 30 ÷ 4 = 7.5. Stretch check: mean 10, distances 1, 0, 0, 1, MAD 0.5.
National Institute of Standards and Technology, NIST/SEMATECH e-Handbook of Statistical Methods, Measures of Scale · U.S. government work, public domain
NIST’s statistics handbook says that describing the spread, or variability, of a data set is one of the most basic jobs in statistics. It asks two questions: how spread out are the values near the center, and how spread out are the tails, the values far from the center?
The handbook compares several measures. The range is the largest value minus the smallest. It uses only the two extreme values, so it says nothing about the spread near the center.
The interquartile range is the 75th percentile minus the 25th percentile. It uses only the middle part of the data.
The handbook’s average absolute deviation is the measure we call the mean absolute deviation. Find the mean, find how far each value is from the mean (ignoring whether it is above or below), and take the average of those distances. The handbook says this measure is less affected by extreme values than some other measures scientists use.
The handbook also shows a histogram as a good way to see both kinds of spread. Its main point is that different measures pay attention to different parts of the data, so the choice depends on what you want to show.
Source: National Institute of Standards and Technology, NIST/SEMATECH e-Handbook of Statistical Methods, Measures of Scale (itl.nist.gov)