Measures of Dispersion

Dr Oluwatoni Adeyemi
Consultant Public Health Physician, LASUTH
Associate Lecturer, LASUCOM
Department of Community Health and Primary Health Care
 Learning Objectives

• Describe the following measures of spread:

– Range, Interquartile range
– Variance, Standard deviation, coefficient of variation
• Match each measure of spread to the measure of central
location with which it is used
• Describe which measures should be used with which types of
data
 What is Dispersion?

• Dispersion refers to the spread or variability of a set of data

values.
• It answers the question: "How much do the values deviate
from the central point (like the mean)?
• In medical research, understanding the variability of data is
essential for evaluating risk, treatment efficacy, and
generalizing findings.
 Central Tendency vs. Dispersion
• Central Tendency (mean, median, mode) provides a summary
of a data set by identifying a central point.
• Dispersion quantifies the degree to which data points vary
from this central point.
 Examples of Variability

Less Variability Higher Variability

Same central location

but …

différent spread
 Measure of Spread
Measure that quantifies the variation or dispersion or spread of
a set of data from its central location

• Also known as:

– “Measure of dispersion”
– “Measure of variation”

• Common measures of spread

– Range
– Interquartile range
– Variance / standard deviation
 Range
Definition:
Range = difference between largest and smallest values

Method for identification

1. Sort data or create frequency distribution
2. Find maximum and minimum values
3a. Subtract minimum from maximum, or
3b. Report minimum and maximum
 Range

Minimum Maximum

Range
• The range is the simplest measure of dispersion, calculated by
subtracting the smallest value in the dataset from the largest.
• It is usually used with the median.

Range = Maximum Value−Minimum Value

Example: Consider a sample of systolic blood pressure readings: 120,

125, 130, 145, 150. What is the range?
Range = 150 - 120 = 30.

Limitations: The range is heavily affected by outliers (extreme values),

making it less robust compared to other measures of dispersion.
 Quartiles
Method for identification
1. Arrange observations in order
2. Use median to divide data into two halves.
3. Exclude median.
Q1 = median of lower half of data
Q3 = median of upper half of data

Method 2 for identification

4. Arrange observations in order
Q1 = value at position (n+1) x (1/4)
Q3 = value at position (n+1) x (3/4)
 Quartiles
• Value divides ordered observations into 4 equal parts.
• 1st quartile is value below which 1/4 of the observations lie.
• 1st quartile equivalent to 25th percentile.
• 2nd quartile equivalent to median or 50th percentile.
• 3rd quartile value above which 3/4 of the ordered observations is located.
Equivalent to 75th percentiles
 Interquartile Range
• The IQR measures the range of the middle 50% of the data,
offering a more robust measure of spread. It is calculated as the
difference between the 75th percentile (Q3) and the 25th
percentile (Q1).
• Difference between 3rd quartile and Ist quartile.
• Concentration on the middle 50% of the ordered observations.
• Not affected by outliers.
• It is often used with the median.
 IQR=Q3−Q1

Example: For the data set 12, 15, 18, 20, 25, 30, 35, 40, get the IQR

Q1 = 15, Q3 = 30 → IQR = 30 - 15 = 15.

Interpretation:
• A higher IQR indicates greater variability in the middle 50% of the data.
• Used in clinical contexts to understand the spread of patient values (e.g.,
drug concentrations in the blood, symptom scores).

• Advantage: Less influenced by extreme values compared to the range.

Obs Finding Q1 and Q3 of Sorted Series,
Age
1
2
27
28
Odd number of values (n = 19)
3 28 Method 2 Steps
4 29
1. Sort (done)
5 29
6 29
2. Q1 = value at position (n+1) x (1/4)
7 30
(19+1) x (1/4)
8 30 5
9 30 Q1 = value at position 5 = 29
10 30
11 30
Q3 = value at position (n+1) x (3/4)
12 31
13 31
(19+1) x (3/4)
14 32 15
15 33 Q3 = value at position 15 = 33
16 34
17 35
18 36
19 37
 Exercise
1. Find Q1, Q2 and Q3 of the following observations:
a. 6, 7, 15, 36, 39, 40, 41, 42, 43, 47, 49
b. 7, 15, 36, 39, 40, 41

2. What’s the interquartile range?

 Variance and Standard Deviation

• The variance and standard deviation are the measures of spread used
most commonly with the arithmetic mean.
• Earlier, the centering property of the mean was described — subtracting
the mean from each observation and then summing the differences adds
to 0.
• This concept of subtracting the mean from each observation is the basis
for the standard deviation.
• To derive the standard deviation, the difference between the mean and
each observation is squared in order to eliminate negative numbers.
• By squaring the difference and then averaging, you can assess the spread
or dispersion around the mean.
 Variance and Standard Deviation
Mean
Measure that quantifies how far away,
on average, each observation is from
the center of the distribution
 Steps for Calculating
Sample Variance and Standard Deviation
1. Calculate the arithmetic mean
2. Subtract the mean from each observed value
3. Square each difference
4. Sum the squared differences (“sum of squares”)
5. Divide the sum of the squared differences by n or n – 1 (= variance)
6. Take the square root of the variance
(= standard deviation)
 Equations for
Sample Variance and Standard Deviation

( xi  x ) 2
_
x : mean s 
2

xi : value
n : number
n 1
s²: variance
s : standard deviation

( xi  x ) 2

s
n 1
 Steps for Calculating
Sample Variance and Standard Deviation
1. Calculate the arithmetic mean
2. Subtract the mean from each observed value
3. Square each difference
4. Sum the squared differences (“sum of squares”)
5. Divide the sum of the squared differences by n or n –
1 (= variance)
6. Take the square root of the variance
(= standard deviation)
 Calculate
Obs
1
Age
27 Variance,
2
3
28
28
Standard
4
5
29
29 1. Deviation
Calculate mean = 589/19 = 31.0
6 29
2. Subtract mean from each value
7 30
8 30 3. Square each difference
9 30
4. Sum the squared differences
10 30
(“sum of squares”)
11 30
12 31 5. Variance: Divide the sum of
13 31 squares by n – 1
14 32
15 33
16 34 6. Standard deviation: Take the
17 35 square root of the variance
18 36
19 37
Sum: 589
 Calculate
Obs
1
Age
27 Variance,
2
3
28
28
Standard
4
5
29
29 1. Deviation
Calculate mean
6 29
2. Subtract mean from each value
7 30
8 30 3. Square each difference
9 30
4. Sum the squared differences (“sum
10 30
of squares”)
11 30
12 31 5. Variance: Divide the sum of squares
13 31 by n – 1
14 32
15 33
16 34 6. Standard deviation: Take the square
17 35 root of the variance
18 36
19 37
Sum: 589
 Calculate
Obs
1
Age
27
Mean
31
Diff
Variance,
2
3
28
28
31
31
Standard
4
5
29
29
31
31 1. Deviation
Calculate mean
6 29 31
2. Subtract mean from each value
7 30 31
8 30 31 3. Square each difference
9 30 31
4. Sum the squared differences (“sum
10 30 31
of squares”)
11 30 31
12 31 31 5. Variance: Divide the sum of squares
13 31 31 by n – 1
14 32 31
15 33 31
16 34 31 6. Standard deviation: Take the square
17 35 31 root of the variance
18 36 31
19 37 31
Sum: 589
 Calculate
Obs
1
Age
27
Mean
31
Diff
-4 Variance,
2
3
28
28
31
31
-3
-3
Standard
4
5
29
29
31
31
-2
-2 1. Deviation
Calculate mean
6 29 31 -2
2. Subtract mean from each value
7 30 31 -1
8 30 31 -1 3. Square each difference
9 30 31 -1
4. Sum the squared differences (“sum
10 30 31 -1
of squares”)
11 30 31 -1
12 31 31 0 5. Variance: Divide the sum of squares
13 31 31 0 by n – 1
14 32 31 1
15 33 31 2
16 34 31 3 6. Standard deviation: Take the square
17 35 31 4 root of the variance
18 36 31 5
19 37 31 6
Sum: 589
 Calculate
Obs
1
Age
27
Mean
31
Diff
-4 Variance,
2
3
28
28
31
31
-3
-3
Standard
4
5
29
29
31
31
-2
-2 1. Deviation
Calculate mean
6 29 31 -2
2. Subtract mean from each value
7 30 31 -1
8 30 31 -1 3. Square each difference
9 30 31 -1
4. Sum the squared differences
10 30 31 -1
(“sum of squares”)
11 30 31 -1
12 31 31 0 5. Variance: Divide the sum of
13 31 31 0 squares by n – 1
14 32 31 1
15 33 31 2
16 34 31 3 6. Standard deviation: Take the
17 35 31 4 square root of the variance
18 36 31 5
19 37 31 6
Sum: 589
 Calculate
Obs
1
Age
27
Mean
31
Diff
-4
Diff Sq’d
16 Variance,
2
3
28
28
31
31
-3
-3
9
9
Standard
4
5
29
29
31
31
-2
-2
4
4 1. Deviation
Calculate mean
6 29 31 -2 4
2. Subtract mean from each value
7 30 31 -1 1
8 30 31 -1 1 3. Square each difference
9 30 31 -1 1
4. Sum the squared differences
10 30 31 -1 1
(“sum of squares”)
11 30 31 -1 1
12 31 31 0 0 5. Variance: Divide the sum of
13 31 31 0 0 squares by n – 1
14 32 31 1 1
15 33 31 2 4
16 34 31 3 9 6. Standard deviation: Take the
17 35 31 4 16 square root of the variance
18 36 31 5 25
19 37 31 6 36
Sum: 589
 Calculate
Obs
1
Age
27
Mean
31
Diff
-4
Diff Sq’d
16 Variance,
2
3
28
28
31
31
-3
-3
9
9
Standard
4
5
29
29
31
31
-2
-2
4
4 1. Deviation
Calculate mean
6 29 31 -2 4
2. Subtract mean from each value
7 30 31 -1 1
8 30 31 -1 1 3. Square each differences
9 30 31 -1 1
4. Sum the squared differences
10 30 31 -1 1
(“sum of squares”)
11 30 31 -1 1
12 31 31 0 0 5. Variance: Divide the sum of
13 31 31 0 0 squares by n – 1
14 32 31 1 1
15 33 31 2 4
16 34 31 3 9 6. Standard deviation: Take the
17 35 31 4 16 square root of the variance
18 36 31 5 25
19 37 31 6 36
Sum: 589 Sum: 142
 Calculate
Obs
1
Age
27
Mean
31
Diff
-4
Diff Sq’d
16 Variance,
2
3
28
28
31
31
-3
-3
9
9
Standard
4
5
29
29
31
31
-2
-2
4
4 1. Deviation
Calculate mean
6 29 31 -2 4
2. Subtract mean from each value
7 30 31 -1 1
8 30 31 -1 1 3. Square each differences
9 30 31 -1 1
4. Sum the squared differences
10 30 31 -1 1
(“sum of squares”)
11 30 31 -1 1
12 31 31 0 0 5. Variance: Divide the sum of
13 31 31 0 0 squares by n – 1
14 32 31 1 1
Variance = 142 / 18 = 7.889
15 33 31 2 4
16 34 31 3 9 6. Standard deviation: Take the
17 35 31 4 16 square root of the variance
18 36 31 5 25
19 37 31 6 36
Sum: 589 Sum: 142
 Calculate
Obs
1
Age
27
Mean
31
Diff
-4
Diff Sq’d
16 Variance,
2
3
28
28
31
31
-3
-3
9
9
Standard
4
5
29
29
31
31
-2
-2
4
4 1. Deviation
Calculate mean
6 29 31 -2 4
2. Subtract mean from each value
7 30 31 -1 1
8 30 31 -1 1 3. Square each differences
9 30 31 -1 1
4. Sum the squared differences
10 30 31 -1 1
(“sum of squares”)
11 30 31 -1 1
12 31 31 0 0 5. Variance: Divide the sum of
13 31 31 0 0 squares by n – 1
14 32 31 1 1
Variance = 142 / 18 = 7.889
15 33 31 2 4
16 34 31 3 9 6. Standard deviation: Take the
17 35 31 4 16 square root of the variance
18 36 31 5 25
SD =
19 37 31 6 36 7.889 = 2.81
Sum: 589 Sum: 142
 Standard Deviation – Properties / Uses
• Standard deviation usually calculated only when data are more
or less normally distributed (bell shaped curve)
• For normally distributed data,
– 68.3% of the data fall within plus/minus 1 SD
– 95.5% of the data fall within plus/minus 2 SD
– 95.0% of the data fall within plus/minus 1.96 SD
– 99.7% of the data fall within plus/minus 3 SD
 How to Choose a Measure of Dispersion
• Range is useful for quick, rough estimates, but it’s sensitive to
extreme values.
• IQR is robust and works well for skewed distributions or when
outliers are present.
• Variance and SD provide more comprehensive insights into
data variability, especially for normally distributed data.
 How to Choose the Appropriate Measure
of Central Location and Spread
• First, look at the data!
• If single peak, symmetrical, large amount of data
– Mean, standard deviation
• If single peak, skewed, large amount of data
– Median, interquartile range
• If single peak, skewed, small amount of data
– Median, range
• If more than one peak
– Mode, range
 Sample Questions
Arithmetic mean Mode
Geometric meanRange
Interquartile range Standard deviation
Median Variance

Question 1: Which ones are considered…

a. Measures of central location?
b. Measures of spread/Dispersion?
 Answer 1
Arithmetic mean Mode
Geometric mean Range
Interquartile range Standard deviation
Median Variance
 Sample Question 2
Mode Range
Median Mean

Question 2: Which ones are affected by one extreme value?

Sample Question 3

Q5. Match the measures of central location with its

usual measures of spread.

 Mode  Standard deviation

 Median
 Arithmetic mean  Range
 Interquartile range
 Answer

Match the measures of central location with its usual measures of

spread.

 Mode  Standard deviation

 Median
 Arithmetic mean  Range
 Interquartile range
 Applications in Medical Research
1. Clinical Trials:
Measures like the SD and IQR help understand how diverse patient
responses are to a treatment, assisting in interpreting the effectiveness
and safety of a drug or intervention.
2. Diagnostic Test Evaluation:
Understanding the spread of test results (e.g., sensitivity, specificity,
positive/negative predictive value) can guide decision-making about
the reliability of diagnostic tools.
3. Epidemiology:
Dispersion measures help assess how widely diseases or conditions are
distributed in a population, which can inform public health strategies.
 Conclusion
• Dispersion is crucial for interpreting the variability in medical
data, which impacts decision-making, risk assessment, and
treatment efficacy.
• Standard deviation, variance, range, and IQR are tools that,
when used together, provide a comprehensive picture of the
data, helping researchers and clinicians make more informed
decisions.