TOPIC: Descriptive Statistics

Single Variable

I. Numerical data – summary measurements

A. Measures of Location
1. Measures of central tendency
Mean; Median; Mode
2. Quantiles - measures of noncentral tendency
Quartiles; Percentiles
B. Measures of Dispersion
Range; Interquartile range; Variance; Standard Deviation;
Coefficient of Variation
C. Measures of Shape
Skewness
5-number summary
Box-and-whisker
Stem-and-leaf
D. Standardizing Data

II. Categorical data

A. Frequencies (also useful for grouped numerical data)
Frequency Distribution
Percentage Distribution
Cumulative Distribution
Histogram
Polygon
Ogive

B. Charts
Bar chart
Pie chart
Pareto diagram

Lecture Notes: Descriptive Statistics Page 1

 Summary Measures

Measures of Location

Measures of Central Tendency

Mean
Median
Mode

The sample mean is the sum of all the observations divided by the number of
observations:

X1 + X 2 + X 3 + … + X n
X =
n

or
n

∑X
i =1
i
X =
n

where ∑Xi is the same as X1 + X2 + X3 + … + Xn

Example:

1
2
2
4
5
10

X = 24 / 6 = 4.0

Lecture Notes: Descriptive Statistics Page 2

Example:

1
1
1
1
51

X = 55 / 5 = 11.0

Note that the mean is affected by extreme values.

MEDIAN

The median is the middle of the data (after data is arranged in ascending or
descending order); half the observations are less than the median and half are
more than the median. To get the median, we must first rearrange the data into
an ordered array. Generally, we order the data from the lowest value to the
highest value.

The median is the data value such that half of the observations are larger than
it and half are smaller. It is also the 50th percentile (we will be learning about
percentiles).

If n is odd, the median is the middle observation of the ordered array.

If n is even, it is midway between the two central observations.

Lecture Notes: Descriptive Statistics Page 3

EXAMPLE:

0
2
3
5 ←
20
99
100

Median = 5

n=7 Since n is odd, the median is the (n+1)/2 ordered observation, or the 4th
observation.

EXAMPLE n=6

10
20
30
40
50
60

Median = 35

Lecture Notes: Descriptive Statistics Page 4

EXAMPLE: Exam scores

0
0
0 ←
0
100

What is the mean? Suppose the prof lets students know the following grading
policy: Anyone who got the mean or better gets an A for the course; anyone
who got below the mean fails. Are we happy or unhappy?

Note that the mean and median are UNIQUE for a given set of data.

Advantage: the Median is not affected by extreme values. In the previous

example, if you change the 60 to 6,000, the median will still be 35. The mean,
on the other hand will change by a great deal.

Problem: Sometimes it is difficult to order data.

The median has 3 interesting characteristics:

1. The median is not affected by extreme values, only by the number of
observations.
2. Any observation selected at random is just as likely to be greater than
the median as less than the median.
3. Summation of the absolute value of the differences about the median
is a minimum:
∑i =1 X i − Median = minimum
n

Lecture Notes: Descriptive Statistics Page 5

 MODE

The mode is the value of the data that occurs with the greatest frequency.

EXAMPLE:

1 1 1 2 3 4 5

The mode is 1 since it occurs three times. The other values only appear once
in the data set.

EXAMPLE:

5 5 5 6 8 10 10 10

Mode = 5, 10
The modes for this data set are 5 and 10. This is a bi-modal dataset.

Problems:
The mode may not exist.
The mode may not be unique.

Lecture Notes: Descriptive Statistics Page 6

 QUANTILES

Measures of non-central location

Quartiles
Deciles
Percentiles

These are all commonly used quantiles.

QUARTILES split the ordered data into four quarters.

Imagine cutting a chocolate bar into four equal parts… How many cuts would
you make? (yes, 3!)

Q1 – First Quartile – 25% of the observations are smaller than Q1 and

75% of the observations are larger than Q1
Q2 – Second Quartile – 50% of the observations are smaller than Q2 and
50% of the observations are larger than Q2. Same as
the Median. It is also the 50th percentile.
Q3 – Third Quartile – 75% of the observations are smaller than Q3and
25% of the observations are larger than Q3

The quartiles, like the median, either take the value of one of the observations,
or the value halfway between two observations.

If n/4 is an integer, the first quartile (Q1) has the value halfway between
the (n/4)th observation and the next observation.

If n/4 is not an integer, the first quartile has the value of the observation
whose position corresponds to the next highest integer.

Lecture Notes: Descriptive Statistics Page 7

EXAMPLE:

210
220
225 Q1 = 225, the 3rd observation
225
225
Q2 = Median - 230
235
240
250 Q3=250, 3rd observation from the bottom
270
280

Please note that this technique is an approximation, and much easier than
using the formula (algorithm). The Q1 and Q3 values you get from MS Excel
may be slightly different.

EXAMPLE: Computer Sales

Original Ordered
data data
3 0
10 2
2 3
--Q1
5 4
9 5
8 6
--Q2
7 7
12 8
10 9
--Q3
0 10
4 10
6 12

Median = 6.5 Mode = 10

Q1 = 3.5 ||||| 25% smaller; 75% larger
Q3 = 9.5 ||||| 75% smaller; 25% larger
∑X = 76
X = 76/12 = 6.33

Lecture Notes: Descriptive Statistics Page 8

 DECILES, PERCENTILES

Similarly,
there are 9 deciles dividing the distribution into 10 equal portions (tenths).
there are 99 percentiles dividing the distribution into 100 equal portions.

In all these cases, the convention is the same. The point, be it a quartile,
decile, or percentile, takes the value of one of the observations or it has a value
halfway between two adjacent observations. It is never necessary to split the
difference between two observations more finely.

Percentiles are used in analyzing the results of standardized exams. For

instance, a score of 40 on a standardized test might seem like a terrible grade,
but if it is the 99th percentile, don’t worry about telling your parents. 

Q1 = which percentile?
Q2 (the median)?
Q3?

Lecture Notes: Descriptive Statistics Page 9

EXERCISE:

n=16

1
1
2
2
2
2
3
3
4
4
5
5
6
7
8
10

Lecture Notes: Descriptive Statistics Page 10

SOLUTION:

n=16

1
1
2
2
2
2
3
3
4
4
5
5
6
7
8
10

X = 65/16 = 4.06
Median = 3.5
Mode = 2
Q1 = 2
Q3 = 5.5

Lecture Notes: Descriptive Statistics Page 11

EXERCISE: # absences
n=16

0
5
3
2
1
0
2
4
3
2
1
0
0
0
6
8

Lecture Notes: Descriptive Statistics Page 12

SOLUTION: # absences
n=16

0
5
3
2
1
0
2
4
3
2
1
0
0
0
6
8

Ordered list: 0 0 0 0 | 0 1 1 2 | 2 2 3 3 | 4 5 6 8

X = 37/16 = 2.31
Median = 2
Mode = 0
Q1 = 0
Q3 = 3.5

Lecture Notes: Descriptive Statistics Page 13

EXERCISES – COMPUTE MEAN, MEDIAN, MODE, QUARTILES

EXERCISE: Reading Level of n=16 students in 8th grade

5
6
6
6
5
8
7
7
7
8
10
9
9
9
9
9

ANS 1.
First, order the data, and split into fourths using the three quartiles, Q1, Q2, and
Q3:

5 5 6 6 || 6 7 7 7 || 8 8 9 9 || 9 9 9 10
Q1 Q2 Q3

Median = Q2 = 7.5
Q1 = 6, Q3 = 9
Mode = 9
Sum = 120
th
X = 120 / 16 = 7.5 average reading level in the 8 grade class

Alternate method: This data set can also be set up to be analyzed as grouped
data. We don’t really have to order the entire data set in this case.

Lecture Notes: Descriptive Statistics Page 14

Original data:
5
6
6
6
5
8
7
7
7
8
10
9
9
9
9
9

Xi (reading level) frequency, fi

5 2
6 3
7 3
8 2
9 5
10 1

Note that ∑fi= n= 16

Lecture Notes: Descriptive Statistics Page 15

To get the mean, median, etc.:

Xi fi (Xi)(fi)
5 2 10
6 3 18
7 3 21
8 2 16
9 5 45
10 1 10
16 (= n) 120 (= sum of ungrouped data)

When data is grouped, how do we get the mean? Median? Mode? Quartiles?

Grouped data can also be presented in a bar chart (later).

EXERCISE: #colds / year

n=16

0 1 2 2 3 3 3 4 4 4 4 5 5 6 8 10

EXERCISE: avg. wait in minutes for a train

n=15

0 6 10 16
4 8 11 17
5 9 12 45
5 10 15

Lecture Notes: Descriptive Statistics Page 16

 MEASURES OF DISPERSION

Dispersion is the amount of spread, or variability, in a set of data.

Why do we need to look at measures of dispersion?

Example:
A company is buying computer chips. These chips must have an average life
of 10 years. The company has a choice of two suppliers. Whose chips should
they buy? They take a sample of 10 chips from each of the suppliers.

Supplier Supplier
A B
(n = 10) (n = 10)
11 years 170 years
11 1
10 1
10 160
11 2
11 150
11 150
11 170
10 2
12 140

XA = 10.8 years XB = 94.6 years

Median = 11 years Median = 145 years
sA = 0.63 years sB = 80.6 years
Range = 2 years Range = 169 years

Supplier B’s chips have a longer average life. However, note that with a 3-year
warranty, supplier A will have no returns while supplier B will have 4/10 or
40% returns.

Lecture Notes: Descriptive Statistics Page 17

There are 5 major measures of dispersion:

Range
Interquartile Range
Standard Deviation
Variance
Coefficient of Variation

RANGE

Range = Largest Value – Smallest Value

Example:

1 2 3 4 8 Range = 8 – 1 = 7

Problem: The range is influenced by extreme values at either end.

INTERQUARTILE RANGE

IQR = Q3 – Q1

From previous example:

210
220
rd
225 Q1 = 225, the 3 observation
225
225
Q2 = Median - 230
235
240
rd
250 Q3=250, 3 observation from bottom
270
280

Lecture Notes: Descriptive Statistics Page 18

Interquartile Range = 250 – 225 = 25

It is basically the range encompassed by the central 50% of the observations in

the distribution.

Problem: The interquartile range does not take into account the variability of
the total data (only the central 50%). We are “throwing out” half of the data.

STANDARD DEVIATION

The standard deviation measures the “average” deviation about the mean. It is
not really the “average” deviation, even though we may think of it that way.

Note: If you take a simple mean, and then add up the deviations about the
mean, i.e., ∑(X- X ), this sum is always going to be equal to 0. Therefore, a
measure of “average deviation” will not work.

∑(X − X )
i =1
2

s= “definitional formula”
n −1

By squaring the deviation, we have ∑(X- X )2 which is a minimum (called the

“least squares property”). No other value subtracted from X and squared will
result in a smaller sum of the deviation squared.

The standard deviation has lots of nice properties, which we will study shortly.

Lecture Notes: Descriptive Statistics Page 19

Example:
Xi Yi
1 0
2 0
3 0
4 5
5 10

Find X , Y , sx, sy

X X (X- X ) (X- X )2
1 3 -2 4
2 3 -1 1
3 3 0 0
4 3 1 1
5 3 2 4
∑=0 10

SX = √10/4 = 1.58

Y Y (Y- Y ) (Y- Y )2
0 3 -3 9
0 3 -3 9
0 3 -3 9
5 3 2 4
10 3 7 49
∑=0 80

SY = √80/4 = 4.47

[CHECK WITH CALCULATOR]

Lecture Notes: Descriptive Statistics Page 20

Note that the population standard deviation is:

∑ ( X − µ) 2

σ= i =1

but it is very rare that we ever take a census of the population and deal with N,

Normally, we work with a sample and calculate the sample measures, like the
sample mean and the sample standard deviation, s:

∑(X − X )
i =1
2

s=
n −1

The reason we divide by n-1 instead of n is to assure that s is an unbiased

estimator of σ. We have taken a shortcut: in the second formula, we are using
X , a statistic, in lieu of μ, a parameter. To correct for this – which has a
tendency to understate the true standard deviation – we divide by n-1 which
will increase s somewhat and make it an unbiased estimator of σ. Later on in
the course we will refer to this as “losing one degree of freedom.”

Lecture Notes: Descriptive Statistics Page 21

 VARIANCE

The variance is
n

∑(X − X ) 2

s2 = i =1

n −1

or, s = Variance

Computational formula. This is what your calculator uses:

n

n
(∑ X i ) 2
∑X − i =1
2
i
n
s2 = i =1

n −1

n
(∑ X i ) 2
∑X − i =1
2
i
n
s = √s2 = i =1

n −1

Lecture Notes: Descriptive Statistics Page 22

EXERCISE: X = # minutes waiting for bus

X
0 5 10 4 8 6 9 0 2 6

EXERCISE: X = hours to complete task

X
6 4 10 4 7 5 9 11

EXERCISE: X = # cups of coffee students drink in a day

X
10 6 9 7 8 0 0 4 5

Lecture Notes: Descriptive Statistics Page 23

 5. COEFFICIENT OF VARIATION

The problem with s2 and s is that they are in the “original” units.

This makes it difficult to compare the variability of two data sets, if they are in
different units OR if the magnitude of the numbers is very different.

Suppose you wish to compare two stocks and one is in dollars and the other is
in yen; if you want to know which one is more volatile, you should use the
coefficient of variation. It is also not appropriate to compare two stocks of
vastly different prices even if both are in the same units. The standard
deviation for a stock that sells for around $300 is going to be very different
than one where the price is around $0.25. The coefficient of variation will be a
better measure of dispersion when comparing the two stocks than the standard
deviation (see example below).
s
CV = x 100%
X

Back to previous example:

Xi Yi
1 0
2 0
3 0
4 5
5 10
X =3 Y =3
s=1.5 s=4.4
8 7

1.58
CVX = x 100% = 52.7%
3

4.47
CVY = x 100% = 149%
3

These two datasets aren’t really that different. Here, s pretty much does the
job.
Lecture Notes: Descriptive Statistics Page 24
Example:
Which stock price is more volatile?

Closing prices over the last 8 months:

Stock Stock B
A
JAN $1.00 $180
FEB 1.50 175
MAR 1.90 182
APR .60 186
MAY 3.00 188
JUN .40 190
JUL 5.00 200
AUG .20 210

Mean $1.70 $188.88

s2 2.61 128.41
s $1.62 $11.33

The standard deviation of B is higher than for A, but A is more volatile:

$1.62
CVA = x 100% = 95.3%
$1.70

$11.33
CVB = x 100% = 6.0%
$188.88

Lecture Notes: Descriptive Statistics Page 25

Exercise: Test Scores
X
0 0 40 50 50 60 70 90 100 100

Compute the mean, median, mode, quartiles (Q1, Q2, Q3), range, interquartile
range, variance, standard deviation, coefficient of variation.

Exercise: Test Scores

X
0 0 40 50 50 || 60 70 90 100 100
Q1 Q2 Q3

Sum = 560
X = 560 / 10 = 56.0
median = Q2 = 55
Q1 = 40, Q3 = 90
[Note: Excel gives these as Q1 = 42.5, Q3 = 85]
mode = 0, 50, 100
Range = 100 – 0 = 100
IQ Range = 90 – 40 = 50
s2 = 11,840 / 9 = 1315.5
s = √1315.5 = 36.27
CV = (36.27 / 56) x 100% = 64.8%

Exercise: # employee absences

X
0 0 1 1 1 2 2 2 2 3 4 6

Exercise: Quiz scores 

X
0 5 10
0 6 10
0 7 10
0 8
0 9
Lecture Notes: Descriptive Statistics Page 26
C. Measures of Shape

A third important property of data is its shape.

1. Central tendency
2. Dispersion
3. Shape

Shape can be described by degree of asymmetry (i.e., skewness).

mean > median positive or right-skewness

mean = median symmetry or zero-skewness
mean < median negative or left-skewness

Positive skewness arises when the mean is increased by some unusually high
values. Negative skewness occurs when the mean is decreased by some
unusually low values.

Example: # hours to complete a task

2 3 8 8 9 10 10 12 15 18 22 63
This guy took a VERY
long time!
Q1 Median Q3

X = 15 – skewed right! Median = 10.

s2 = 2868 / 11 = 260.79
s = 16.25
CV = 107.7%

Lecture Notes: Descriptive Statistics Page 27

To measure skewness, MS Excel uses the Pearson coefficient of skewness:
3( X − Median)
SkewnessP =
s
This formula measures departure from symmetry. If the data is symmetrically
distributed, this means that mean=median=mode.

We do not have to remember these formulas. We can get the skewness

measures as output from MS Excel.

NOTE: An alternate formula used to measure skewness, if the mode exists, is:
X − Mode
SkewnessP = or, if Mode is not known:
s

We do not have to know these formulas. We get the skewness measures as

output from MS Excel.

Lecture Notes: Descriptive Statistics Page 28

EXAMPLE: # defects in a sample of 12 cars
(n=12)

2 3 8 8 9 10 10 12 15 18 22 63

Lecture Notes: Descriptive Statistics Page 29

 Descriptive Statistics in MS Excel

This is a good point at which to make some time to get more proficient in MS Excel. See the
class handout “Obtaining Descriptive Statistics in MS Excel” and follow the instructions to
do the problems in the lecture notes and in the Homework Assignment.

Lecture Notes: Descriptive Statistics Page 30

 STANDARDIZING DATA

Z-scores: We can convert the original scores to new scores with

X̄ = 0 and s = 1.
[Note: you have a pure number with no units of measurement.]

Any score below the mean will now be negative.

Any score at the mean will be 0.
Any score above the mean will now be positive.

Example:
X Z
0−5
0 -1.34
3.74

2−5
2 -.80
3.74

4−5
4 -.27
3.74

6−5
6 .27
3.74

8−5
8 .80
3.74

10 − 5
10 1.34
3.74

Lecture Notes: Descriptive Statistics Page 31

Example:

X Z X Z X Z

65 -0.45 65 -0.81 65 -1.40

73 -0.11 73 -0.38 73 -0.79
78 0.10 78 -0.10 78 -0.40
69 -0.28 69 -0.60 69 -1.09
78 0.10 78 -0.10 78 -0.40
7 -2.89 <= 97 0.94 97 1.07
23 -2.21 23 -3.12 <= 93 0.76
98 0.94 98 0.99 98 1.14
99 0.99 99 1.05 99 1.22
99 0.99 99 1.05 99 1.22
97 0.90 97 0.94 97 1.07
99 0.99 99 1.05 99 1.22
75 -0.02 75 -0.27 75 -0.63
79 0.14 79 -0.05 79 -0.32
85 0.40 85 0.28 85 0.14
63 -0.53 63 -0.92 63 -1.56
67 -0.36 67 -0.70 67 -1.25
72 -0.15 72 -0.43 72 -0.86
73 -0.11 73 -0.38 73 -0.79
93 0.73 93 0.72 93 0.76
95 0.82 95 0.83 . 95 0.91

Mean 75.57 Mean 79.86 Mean 83.19

Std. Std. Std.
Dev. 23.75 Dev. 18.24 Dev. 12.96

For standardized data, if it is normally distributed, 95% of the data will be

between 2 standard deviations about the mean.

If the data follows a normal distribution, 95% of the data will be between -1.96
and +1.96. 99.7% of the data will fall between -3 and +3. 99.99% of the data
will fall between -4 and +4.

No matter what you are measuring, a Z-score of more than +5 or less than – 5
would indicate a very, very unusual score.

Worst case scenario: 75% of the data are between 2 standard deviations about
the mean. [Chebychev.]
Lecture Notes: Descriptive Statistics Page 32
 5-NUMBER SUMMARY
[Developed by Tukey]

Median
Q1 Q3
Lowest Highest

Example:
2 3 8 8 9 10 10 12 15 18 22 63

Q1 Median Q3

X = 15
s2 = 2868 / 11 = 260.79
s = 16.25
CV = 107.7%

10
8 16.5
2 63

This data is right-skewed.

In right-skewed distributions, the distance from Q3 to Xlargest is significantly

greater than the distance from Xsmallest to Q1.

Also, midrange > midhinge > median

With left-skewed data,

(Xsmallest to Q1) > (Q3 to Xlargest)
Median > midhinge > midrange

Lecture Notes: Descriptive Statistics Page 33

 BOXPLOT

Vertical line drawn within the box = median

Vertical line at the left side of box = Q1
Vertical line at the right side of box = Q3
Dashed line on left connects left side of box with Xsmallest (lower 25% of data)
Dashed line on right connects right side of box with Xlargest (upper 25% of data)

EXAMPLE:
For the data set above, the Box-and-Whisker Plot would look like this:

(Right-skew)

A “bell-shaped” data distribution would look like this:

(Symmetric)

Lecture Notes: Descriptive Statistics Page 34

 FREQUENCY DISTRIBUTION

Frequency Distribution: Records data grouped into classes and the number of
observations that fell into each class.

A frequency distribution can be used for: categorical data; for numerical data
that can be grouped into categories (or, classes); and for numerical data with
repeated observations.

A Percentage Distribution records the percent of the observations that fell into
each class.

Example: A (fictitious) sample was taken of 200 professors at a local college.

Each was asked for his or her weekly salary. The responses ranged from about
$520 to $590. If we wanted to display the data in, say, 7 equal intervals, we
would use an interval width of $10.

width of interval = range/number of classes

= $70/7 = $10/class.

FREQUENCY DISTRIBUTION
Weekly earnings Frequency Percentage
520 and under 530 6 3 %
530 " " 540 30 15
540 " " 550 38 19
550 " " 560 52 26
560 " " 570 42 21
570 " " 580 24 12
580 to 590 8 4
n = 200 100 %

Lecture Notes: Descriptive Statistics Page 35

A Cumulative Distribution focuses on the number or percentage of cases that
lie below or above specified values rather than within intervals.

Frequency Percentage
less than 520 0 0 %
" " 530 6 3
" " 540 36 18
" " 550 74 37
" " 560 126 63
" " 570 168 84
" " 580 192 96
" " 590 200 100

Lecture Notes: Descriptive Statistics Page 36

 FREQUENCY HISTOGRAM

FREQUENCY POLYGON

Lecture Notes: Descriptive Statistics Page 37

 CUMULATIVE FREQUENCY DISTRIBUTION

CUMULATIVE PERCENTAGE DISTRIBUTION

Lecture Notes: Descriptive Statistics Page 38

Other types of charts for data presentation, categorical variable:

Bar chart

Pie chart

Pareto Diagram

Lecture Notes: Descriptive Statistics Page 39

 TOPIC: Descriptive Statistics
Two Variables

I. Categorical data

A. Contingency Tables

B. Side-by-Side Bar Chart

II. Numerical data – i.e., looking for relationships in bivariate data

A. Scatter Plot (Ch. 2)

B. Correlation (Ch. 3)

C. The Regression Line (Ch. 12)

Lecture Notes: Descriptive Statistics Page 40

 A. The Contingency Table

Two categorical variables are most easily displayed in a contingency table.

This is a table of two-way frequencies.

Example: “Who would you vote for in the next election?”

Mal Female
e
Republican 250 250 500
Candidate
Democrat Candidate 150 350 500
400 600 1000

Also for two-way percentages:

Mal Female
e
Republican 25% 25% 50%
Candidate
Democrat Candidate 15% 35% 50%
40% 60% 100%

Lecture Notes: Descriptive Statistics Page 41

 B. The Side – by – Side Bar Chart

Relative Performance (Source: [Link])

Lecture Notes: Descriptive Statistics Page 42

 What do we do with 2 numerical variables? We can graph them:

Scatter Plot

Example:
Y (Grade) 100 95 90 80 70 65 60 40 30 20
X 73 79 62 69 74 77 81 63 68 74
(Height)
in inches

r = .12
r2 = .01

Lecture Notes: Descriptive Statistics Page 43

For more about Scatter Plots, see the Correlation lecture notes.

Lecture Notes: Descriptive Statistics Page 44