4.3 Measures of central tendency and dispersion
Studying this course? Keep your progress in one place.
Key facts
- Mean ; the median is the middle of the ordered data; the mode is the most frequent value.
- Grouped data: use mid-interval values, and the result is an estimate. The modal class has the highest frequency.
- IQR ; and come from technology.
- For : the mean becomes and the standard deviation becomes .
- An extreme value distorts the mean and standard deviation, but barely moves the median and IQR.
Mean, median and mode
Add and divide for the mean, order the data for the median, and count for the mode.
For a frequency table the mean is . At SL the data set is treated as the whole population, so the formula booklet writes the mean as .
The median is the th value once the data are in order; with an even number of values take the mean of the middle two. The mode is the most frequent value; there can be several, or none.
If you know the mean you know the total, . That is the quickest way to handle a value being added or removed.
- Mean
- Frequency table
- Medianthe th ordered value
Worked example
Find the mean, median and mode of 3, 5, 5, 8, 9.
- 1
The data are in order. Mean: .
- 2
Median: there are 5 values, so take the 3rd value.
- 3
Mode: 5 appears twice, more than any other value.
The mean of 9 values is 7. A tenth value of 17 is added. What is the new mean?
Grouped data
Replace each class by its mid-interval value and use .
When data are grouped into classes such as , the exact values are lost. To estimate the mean, replace every value in a class by its mid-interval value (here 27.5) and use .
The modal class is the class with the highest frequency. On this course it is only used with equal class widths.
The answer is an estimate because it assumes the values in each class are centred on the mid-interval value.
Worked example
Estimate the mean of: (frequency 4), (6), (10). State the modal class.
- 1
Mid-interval values: 5, 15, 25.
- 2
.
- 3
, so the mean is .
- 4
The highest frequency is 10, in the class .
Why is the mean from grouped data only an estimate?
Measures of dispersion
Range uses two values, the IQR uses the middle half, and standard deviation uses every value.
Dispersion measures how spread out the data are.
- Range largest smallest: depends only on the two most extreme values.
- Interquartile range : the spread of the middle 50%, so extreme values do not affect it.
- Standard deviation : a typical distance from the mean. The variance is .
Find and the variance with technology only. At SL use the population standard deviation, labelled on most calculators, not . The standard deviation has the units of the data; the variance has squared units.
| Range | IQR | Standard deviation | |
|---|---|---|---|
| Uses | Two extreme values | Middle 50% | Every value |
| Affected by extremes | Yes | No | Yes |
| Found by | Largest − smallest | Q₃ − Q₁ | Technology |
Range
- Uses:
- Two extreme values
- Affected by extremes:
- Yes
- Found by:
- Largest − smallest
IQR
- Uses:
- Middle 50%
- Affected by extremes:
- No
- Found by:
- Q₃ − Q₁
Standard deviation
- Uses:
- Every value
- Affected by extremes:
- Yes
- Found by:
- Technology
Which measure of spread is not affected by one extreme value?
Quartiles and technology
Use technology; the quartiles split the ordered data into quarters, though methods differ slightly.
The lower quartile and upper quartile split the ordered data into quarters. For discrete data you find them with technology.
Different methods exist: some calculators exclude the median and take the median of each half, others interpolate. For 38, 39, 41, 42, 44, 45, 46, 47, 48, 51, 120, one method gives and , another and . IB markschemes accept values from standard methods.
The value 120 is far from the rest, yet the quartiles barely notice it.
Why might your quartiles differ slightly from a markscheme's?
Effect of constant changes
Adding a constant shifts the mean but not the spread; multiplying scales both.
If every value is transformed by , adding shifts the data but does not change the spread, while multiplying by is scaling. The median and mode transform like the mean.
- x
- y = 2x + 3
- Mean
- Standard deviation
- Variance
Worked example
Temperatures have mean and . Find the mean and in degrees Fahrenheit, using .
- 1
Here and .
- 2
Mean: .
- 3
Standard deviation: . The has no effect.
Data have mean 10 and standard deviation 2. Each value is changed by . What are the new mean and standard deviation?
Choosing measures
For skewed data or an extreme value use the median and IQR; for roughly symmetric data use the mean and standard deviation.
An extreme value pulls the mean and the standard deviation towards it but barely moves the median and IQR.
For skewed data, or data with an extreme value, the median and IQR usually describe a typical value and the spread better. For roughly symmetric data the mean and standard deviation use all of the data.
- Without 40
- With 40
| Roughly symmetric data | Skewed, or an extreme value | |
|---|---|---|
| Centre | Mean | Median |
| Spread | Standard deviation | IQR |
Roughly symmetric data
- Centre:
- Mean
- Spread:
- Standard deviation
Skewed, or an extreme value
- Centre:
- Median
- Spread:
- IQR
A data set has one very large extreme value. Which pair best describes it?
Try an exam question
(a) The mean of 9 values is 7. A tenth value of 17 is added. Find the new mean. (b) Another data set has mean 10 and standard deviation 2. Each value is changed using . Find the new mean and standard deviation.
[4 marks]
- [1]Old total , so the new total is .
- [1]New mean .
- [1]New mean .
- [1]New standard deviation .
That's the notes covered.
Carry on to the next subtopic.
Try a question
Find the mean number of books read.
1 mark, marked against the mark scheme
Find the mean number of books read.
1 mark, marked against the mark scheme