7.2 Measures of Central Tendency (Mean, Median, Mode) and Spread
Key Takeaways
A statistical question anticipates variability. Describe what was measured, how many observations there were and the units.
A mean uses all numerical values. A median uses their order and is less sensitive to extreme magnitudes. A mode describes the most frequent value or category.
Range is the full maximum-to-minimum span. IQR measures the distance between Q1 and Q3. Mean absolute deviation averages the distances from the mean.
Use the question's quartile convention. A box plot shows quartile positions without guaranteeing exact counts in individual parts.
Find out whether whiskers reach the actual extremes or only the outermost unflagged values. Include flagged values when finding the full range.
7.2 Measures of Central Tendency (Mean, Median, Mode) and Spread
Quick Answer: Use the mean, median or mode to summarize center or frequency. Use range, interquartile range or mean absolute deviation to describe spread. A box runs from the first to the third quartile and marks the median. Check the quartile calculation and whisker convention before comparing plots or flagging an unusual value.
The Three Measures of Central Tendency
A statistical question anticipates answers that vary. “How long do visitors wait for a room reservation at this center?” calls for multiple wait times. “How long did one named visitor wait today?” asks for a single value. Describe a set of wait times by identifying what was measured, the sample size, the units, the shape, a typical value and the spread. A center and a spread answer different questions.
1. The Arithmetic Mean
The mean (commonly called the average, denoted x̄ for a sample or μ for a population) is calculated by summing all individual numerical values and dividing that sum by the total number of observations (n):
Finding a Missing Score for a Target Average (Working Backward)
For a target arithmetic mean of equally weighted scores, work backward from the total needed.
- Step 1: Calculate the Total Required Points: Multiply the desired target average by the total number of trials including the final one: Required Total = Target Mean × n_total.
- Step 2: Sum the Completed Scores: Current Sum = x1 + x2 + ... + x_(n-1).
- Step 3: Subtract to Find the Missing Score:
Example: In a five-round practice exercise with equal weights, a learner scores 82, 88, 79 and 91 in the first four rounds. What fifth score gives a mean of 86?
- Total required points: 5 × 86 = 430 points.
- Sum of current scores: 82 + 88 + 79 + 91 = 340 points.
- Required 5th score: 430 - 340 = 90 points.
2. The Median
The median is the middle value of ordered data, or the average of the two middle values for an even number of observations. Ties can leave some values equal to the median.
Critical Rule: You must always arrange the data in numerical order (from least to greatest) before locating the median. Calculating the median on unsorted data is an immediate error.
- Case 1: Odd Number of Values (n is odd): The median is the single middle value located at position (n + 1) / 2. Example: For dataset {12, 15, 18, 22, 29} (n = 5), the median is the 3rd value: 18.
- Case 2: Even Number of Values (n is even): There is no single middle value. The median is the arithmetic mean of the two central values located at positions n / 2 and (n / 2) + 1: Example: For dataset {14, 19, 21, 25, 28, 34} (n = 6), the two central values are 21 and 25. The median is (21 + 25) / 2 = 46 / 2 = 23.
3. The Mode
The mode is the value that appears with the highest frequency in a dataset:
- Unimodal: A dataset with exactly one value that occurs most often (e.g., {4, 7, 7, 7, 9, 12}, mode = 7).
- Bimodal: A dataset with two distinct values that tie for the highest frequency (e.g., {3, 5, 5, 8, 9, 9, 11}, modes = 5 and 9).
- Multimodal: Several values can share the largest frequency. The earlier example has two modes, and a set can have more than two.
- No distinct mode: If all distinct values occur equally often, none occurs more often than the others. In problems that use a no-mode convention, report no mode, for example for .
- Nominal categories: Use the mode for categories that have no natural order, such as choices of paint color. An ordered category such as low, medium or high can also have a median category.
Selecting the Best Measure: Outliers and Skewness
Choose a summary that answers the question about the data. The distribution shape and any unusually large or small values can affect that choice.
Outlier Sensitivity vs. Resistance (Robustness)
- Mean is Sensitive: Because the formula sums every data point, a very extreme value can pull the mean toward it. For example, in a small company where nine workers earn $40,000 and the CEO earns $1,000,000, the mean salary is $136,000—an accurate per-person average that is higher than nine of the ten individual salaries.
- Median is Resistant (Robust): Changing the magnitude of the largest salary in this example leaves the middle positions unchanged, so the median is $40,000. The median is less sensitive to extreme magnitudes, but adding or removing observations, or changing a value enough to move it past the middle, can change it.
Choosing a Summary and Reading Skew
| Measure | What it helps answer | Limitation |
|---|---|---|
| Mean | What is the per-observation average if the total is shared evenly? | Unusually high or low values can shift it substantially. |
| Median | What falls in the middle of the ordered observations? | Values far from the middle contribute by rank, without their full magnitude. |
| Mode | Which value or category occurs most often? | There may be a tie or no distinct mode. |
Skew describes the direction of a longer tail: high values for right skew, low values for left skew. A high extreme often pulls the mean upward, but the exact order of mean, median and mode is not a rule for identifying skew. For example, the symmetric data have both mean and median 3 and modes at 1 and 5. Use the displayed distribution to describe shape and calculate or read the summaries separately.
Measures of Spread and Variability
Central tendency describes where data clusters, but it reveals nothing about how widely dispersed the values are. Two datasets can share an identical mean of 50, yet one may span from 48 to 52 while the other spans from 0 to 100.
1. Range
The range is the simplest measure of dispersion: the arithmetic difference between the highest and lowest observed values: Range = Maximum Value - Minimum Value Limitation: Range reports the full span. A single extreme value can make that span much wider than the interval containing most observations.
2. The Five-Number Summary and Quartiles
List the minimum, first quartile (Q1), median (Q2), third quartile (Q3) and maximum. Q1 and Q3 mark positions near the lower and upper quarters of the ordered data. If values repeat or the number of observations is small, an interval between adjacent quartile values need not contain exactly one fourth of the observations. The five numbers describe center and spread but do not reconstruct the distribution.
Several conventions exist for calculating quartiles. Follow the method specified in the question or used by its data display. In this lesson's median-of-halves method, sort the data and find the overall median. If the count is even, separate the two equal halves. If the count is odd, leave the overall middle observation out of both halves. Then take the median of the lower half for Q1 and the median of the upper half for Q3. This procedure requires at least one observation in each half.
3. The Interquartile Range (IQR)
The interquartile range (IQR) is the span between the first and third quartiles, a measure of spread around the middle:
Changing the magnitude of an extreme value often leaves Q1 and Q3 unchanged. That makes IQR a useful measure of central spread when the range is much wider, although changes in the data or in the calculation convention can change IQR.
4. Mathematical Outlier Detection: The 1.5 × IQR Rule
The 1.5 × IQR rule flags values beyond a pair of fences for this particular method:
- Step 1: Compute IQR: IQR = Q3 - Q1.
- Step 2: Calculate the Step Factor: Multiply IQR by 1.5 (1.5 × IQR).
- Step 3: Calculate the Lower Fence: Values strictly below this fence are flagged.
- Step 4: Calculate the Upper Fence: Values strictly above this fence are flagged. A value equal to a fence stays inside. Check a flagged observation in context before considering any change to the data.
Mean Absolute Deviation
The mean absolute deviation (MAD) from the mean is the average distance of the observations from their mean. Find the mean first. Then subtract it from each value, take each absolute value and find the mean of those distances. The nonnegative distances avoid cancellation and retain the original unit.
For wait times of 8, 10, 10 and 12 minutes, the mean is minutes. The absolute distances are 2, 0, 0 and 2 minutes, so the MAD is minute. These four times were, on average, one minute from the mean. Knowing only that the mean was 10 minutes would not give that information.
Reading and Interpreting Box-and-Whisker Plots
A box plot locates the quartiles on a scaled number line. Draw the box between Q1 and Q3 and mark the median inside it. A wider box means a larger IQR. The widths of the parts measure distances on the scale.
- In a min-to-max box plot, extend the whiskers to the actual minimum and maximum. Their difference gives the range.
- In a modified box plot using the 1.5 × IQR rule, draw the whiskers out to the smallest and largest observations still within the fences, and mark flagged values individually. The fence is a cutoff rather than the whisker endpoint. Use all observations including the flagged extremes when computing the full range.
Worked example: For measured times minutes, the median is 9. Excluding the overall median from the halves gives Q1 = 6.5 and Q3 = 12.5, so IQR = 6 minutes. The fences are and . For the modified plot, draw the box from 6.5 to 12.5 with a line at 9. Extend the whiskers to 4 and 13, and put a separate dot at 23. The full range is minutes. A min-to-max plot of the same data would extend the right whisker all the way to 23.
Comparing Distributions in Context
Use the same time units and scales. A small distance between Q1 and Q3 means that the middle part occupies less space on the scale. For ties and small samples, do not use the box parts to count exact proportions. Long whiskers or a displaced median can suggest asymmetry, but a histogram or line plot shows detail such as gaps and clusters that the five-number summary leaves out.
Suppose two teams measured fulfillment times the same way. One team had a median of 18 minutes and an IQR of 4 minutes, and the other had a median of 24 minutes and an IQR of 4 minutes. The first team's median is 6 minutes lower, a difference of times the common IQR. Identical IQRs show equal central spread by that measure. Individual times can still overlap, and the sample summaries do not settle which team finishes every order faster.
In a practice course with five equally weighted assessments, a learner has scores of 78, 85, 81 and 89 so far. What is the mean of the four scores and the minimum fifth score needed to bring the overall mean to at least 84?
Current mean = 83.00; Minimum score required on 5th exam = 86
Current mean = 83.25; Minimum score required on 5th exam = 85
Current mean = 83.25; Minimum score required on 5th exam = 87
Current mean = 83.50; Minimum score required on 5th exam = 88
A practice data set gives these turnaround times (in hours) for 11 repairs: 14, 18, 19, 21, 22, 24, 25, 27, 28, 31, and 46. Using the median of each half after excluding the overall median, what are the median, first quartile (Q1), third quartile (Q3), and interquartile range (IQR) for this dataset, and does the dataset contain any statistical outliers under the 1.5 × IQR rule?
Median = 24; Q1 = 19; Q3 = 28; IQR = 9; 46 is an outlier because 46 > 41.5
Median = 24; Q1 = 18.5; Q3 = 29.5; IQR = 11; there are no outliers
Median = 22; Q1 = 19; Q3 = 28; IQR = 9; 14 is an outlier because 14 < 15
Median = 24; Q1 = 21; Q3 = 28; IQR = 7; 46 is an outlier because 46 > 38.5
A distribution center compares the order fulfillment times (in minutes) of two warehouse shifts using parallel min-to-max box plots. Shift 1 displays: Minimum = 12, Q1 = 18, Median = 24, Q3 = 30, Maximum = 42. Shift 2 displays: Minimum = 15, Q1 = 22, Median = 28, Q3 = 33, Maximum = 40. Which statement accurately compares the medians, interquartile ranges and full ranges of the two shifts?
Shift 2 has the lower median. Shift 1 has the smaller IQR and the larger range.
Shift 1 has the lower median. Shift 1 also has the smaller IQR and the larger range.
Shift 1 has the lower median. Shift 2 has the smaller IQR, and the shifts have equal ranges.
Shift 1 has the lower median and larger range. Shift 2 has the smaller IQR.
Four recorded waiting times are 4, 6, 10 and 12 minutes. What is the mean absolute deviation from the mean?
0 minutes
3 minutes
4 minutes
8 minutes
Sections you finish are checked off in the contents.