Where Did These Ideas Come From?
People have been collecting data for thousands of years. Ancient farmers counted their sheep. Governments counted their people. But for a long time, nobody had a simple way to describe a whole pile of numbers with just one or two values. That changed when mathematicians developed measures of center (one number that represents the "middle" of a data set) and measures of variability (a number that tells you how spread out the data is).
The key question these thinkers kept running into was: Which summary number should I use? The answer depends on two things — the shape of the data and the context (the real-world story behind it). That's exactly what this lesson is about.
Core Ideas You Need to Know
Before we can choose the best measure, we need to understand what our options are. There are two families of summary numbers: ones that describe the center and ones that describe the spread (variability).
Mean (Average)
Median (Middle Value)
Mean Absolute Deviation (MAD)
Interquartile Range (IQR)
Seeing the Three Shapes of Data
Data sets come in different shapes. The shape tells you a lot about which measure of center to use. Here are the three main shapes you'll see: symmetric, skewed left, and skewed right.
In the symmetric distribution above, the bars make a mirror image — the left side looks the same as the right. Notice how the mean and median are at the exact same spot. When data is symmetric, either measure works great!
In the skewed right distribution, most values are clustered on the left, but a few large values stretch out to the right (like a tail). The mean gets "pulled" toward that long tail, making it higher than the median. The median stays closer to where most of the data actually is.
In the skewed left distribution, it's the opposite — the tail stretches to the left. The mean gets pulled down toward the low values.
The Formulas (and What They Mean)
Let's look at each formula. Don't worry — we'll use easy numbers and take it one step at a time.
Notice the pairs: the mean goes with the MAD, and the median goes with the IQR. You always pick a pair — a center measure and a spread measure that match each other.
Shape and Context: The Decision Guide
Choosing the right measures isn't just about math — it's also about the context (what the data is actually about). Let's see how shape and context work together to guide your choice.
Here's a handy reference table that puts it all together:
| DATA SHAPE | BEST CENTER | BEST SPREAD | REAL-WORLD EXAMPLE |
|---|---|---|---|
| Symmetric, no outliers | Mean | MAD | Heights of all 6th graders in a class |
| Skewed right | Median | IQR | Amount of allowance students get (most get a little, a few get a lot) |
| Skewed left | Median | IQR | Scores on an easy quiz (most score high, a few score very low) |
| Has outliers | Median | IQR | Number of pets at home (most have 0–3, one kid has 15) |
The context matters too. Even if data looks roughly symmetric, if you know there's a good reason for an extreme value (like one student in your survey is a famous YouTuber who makes millions), you might still prefer the median because the mean would be misleading.
Worked Example: Choosing the Right Measures
A teacher asked 9 students how many books they read over the summer. Here are the results:
Mean + MAD vs. Median + IQR: When to Use Each
Let's compare these two pairs side by side so you can always pick the right one.
| FEATURE | MEAN + MAD | MEDIAN + IQR |
|---|---|---|
| Uses every data point? | Yes — every value matters | No — focuses on the middle portion |
| Affected by outliers? | Yes — extreme values pull the mean | No — outliers don't change the median much |
| Best for which shape? | Symmetric data, no outliers | Skewed data or data with outliers |
| Strength | Very precise when data is balanced | Gives a reliable "typical" value even with weird data |
| Weakness | Can be misleading if data is lopsided | Ignores some information (extreme values) |
| Example context | Temperatures in a city over a week | Prices of homes in a neighborhood |
Looking Ahead: Where This Leads
In later grades, you'll learn even more powerful ways to describe data. Here's a sneak peek at how the ideas from this lesson connect to what's coming:
| WHAT YOU KNOW NOW | WHAT'S COMING NEXT |
|---|---|
| MAD (Mean Absolute Deviation) | Standard deviation — a more commonly used measure of spread in high school and beyond |
| Describing shape as "symmetric" or "skewed" | The normal distribution (bell curve) — a specific symmetric shape that shows up everywhere in nature |
| Using context to choose measures | Statistical reasoning — deciding which graphs, tests, and summaries to use based on the situation |
| Median and IQR | Box plots — visual displays that show the median, Q1, Q3, and outliers all in one picture |
The skill you're building right now — thinking carefully about which number best represents your data — is one of the most important skills in all of statistics. Data scientists, doctors, sports analysts, and even video game designers use this kind of thinking every single day.
Practice Problems
8, 10, 11, 12, 13, 14, 65 Find the mean and the median. Which one better represents the "typical" age at the dinner?20, 25, 25, 30, 30, 30, 35, 35, 40 (a) Is this data symmetric, skewed left, or skewed right? (b) Find the mean and median. (c) Which pair of measures (mean + MAD or median + IQR) would you recommend? Why?$150, $180, $200, $210, $220, $240, $1,950 (a) Verify the mean is about $450,000 (remember, values are in thousands). (b) Find the median home price. (c) A family is shopping for a home and expects to find one near the "average" price. Would the mean or median give them a more realistic expectation? Explain using both the data shape and context.Lesson Summary
When you summarize a data set, you have two pairs of tools to choose from. The mean (average) and MAD (mean absolute deviation) work best when data is symmetric and has no outliers, because the mean sits right at the center and the MAD accurately describes how spread out the values are. The median (middle value) and IQR (interquartile range) are the better choice when data is skewed (lopsided) or contains outliers (extreme values), because these measures aren't pulled off-center by unusual data points.
Beyond just looking at the shape, you should always think about the context — what the data is actually about. Situations like home prices, salaries, and scores on very easy or very hard tests often produce skewed data, which means the median and IQR will tell a more honest story. The most important takeaway is this: there is no single "best" measure. The best measure is the one that matches your data's shape and your real-world situation. Always look at your data, check the shape, consider the context, and then choose wisely.