SW388R6

Measures of Variability
Data Analysis
and Computers I

Slide 1

Standard Deviation and Variance

Range

Index of Qualitative Variation

Preferred measure
 SW388R6

Measures of Variability
Data Analysis
and Computers I

Slide 2

 Measures of variability (also called measures of

dispersion) use a single number or data value to
indicate variability, or differences between individual
scores and the measure of central tendency.

 Measures of variability are paired with measures of

central to describe or summarize a distribution. Like
measures of central tendency, the preferred measure
of variability is based on the level of measurement of a
variable.

 Statistically, measures of variability are inidcators of

the errors we making in using a measure of central
tendency as the best estimate of the data value for all
subjects in a distribution.
 SW388R6

The Standard Deviation and Variance - 1

Data Analysis
and Computers I

Slide 3

 The standard deviation is the measure of variability

that is paired with the mean as the measure of central
tendency.

 The difference between each case in a distribution

and the mean is called the deviation. The sum of all
the deviations is equal to zero.

 If we add up the squared deviations for all cases in a

distribution, we produce a measure of error called the
total sum of squares. If we divide the total sum of
squares by the number of cases, we produce the
variance. The square root of the variance is the
standard deviation.
 SW388R6

The Standard Deviation and Variance - 2

Data Analysis
and Computers I

Slide 4

 Standard deviation and variance should only be

computed for interval level variables.

 The standard deviation (and variance) are the

preferred measure of central tendency for an
interval level variable, unless the variable is badly
skewed. The standard deviation is more frequently
used.
 SW388R6

The Range
Data Analysis
and Computers I

Slide 5

 The range is the the difference between the largest

and smallest value in a distribution.

 The range should only be computed for ordinal and

interval level variables.

 The range is the preferred measure of central

tendency for ordinal level variables and interval level
variables that are badly skewed.
 SW388R6

The Index of Qualitative Variation

Data Analysis
and Computers I

Slide 6

 The Index of Qualitative Variation (IQV) is a measure

of the distribution of cases not in the modal
category.

 The IQV may be computed for any level variable. If

the variable is a continuous interval level variable,
values should be grouped before computing the
mode.

 The IQV is the preferred measure of central tendency

for nominal level variables.

 We will not use the IQV because SPSS does not

compute it.
 SW388R6

Request measures of variability

Data Analysis
and Computers I

Slide 7

To compute measures of
central tendency in SPSS,
select the Descriptive
Statistics | Frequencies
command from the Analyze
Other SPSS procedures also menu.
calculate measures of central
tendency, but include a
somewhat different list.

For example, the Descriptives

procedure calculates the mean
and median, but not the mode,
 SW388R6

Select the variable for the analysis

Data Analysis
and Computers I

Slide 8

First, move the

variable “sex” to the
list of Variable(s).

Second, click on the

Statistics… button to
select the specific
measures we want.
 SW388R6

Request the specific measures

Data Analysis
and Computers I

Slide 9

First, mark the

checkboxes for
Standard Deviation,
Variance, and Range.

Since the SPSS output

for this command is Third, click on
not unwieldy, is is the Continue
easiest to request all button to
measures that we complete the
might possibly need. specifications.

Second, mark the

checkbox for Skewness
which we will need for
interval level variables.
 SW388R6

Complete request for measures of variability

Data Analysis
and Computers I

Slide 10

To complete the request for

measures of central
tendency, click on the OK
button.

To reduce the output, you can

clear the Display frequency
tables checkbox and the
frequency distribution will not
be included in the output.
 SW388R6

SPSS output for measures of variability

Data Analysis
and Computers I

Slide 11

The table of Statistics contains

the measures of variability and
skewness that we will need to
solve our homework problems.
 SW388R6

Problem 1
Data Analysis
and Computers I

Slide 12

Based on the dataset GSS2000.SAV, is the following statement

true, false, or an incorrect application of a statistic?

The range for the variable "RS HIGHEST DEGREE" is 4.

1. True
2. False
3. Incorrect application of a statistic
 SW388R6

Solution 1
Data Analysis
and Computers I

Slide 13

The range requires that

the variable be ordinal or
interval. The variable "RS
HIGHEST DEGREE" is
ordinal, satisfying this
requirement.

The range for the variable "RS

HIGHEST DEGREE" is 4.

The answer to the question is true.

 SW388R6

Problem 2
Data Analysis
and Computers I

Slide 14

Based on the dataset GSS2000.SAV, is the following statement

true, false, or an incorrect application of a statistic?

The standard deviation for the variable "RS HIGHEST DEGREE" is

1.16.

1. True
2. False
3. Incorrect application of a statistic
 SW388R6

Solution 2
Data Analysis
and Computers I

Slide 15

The standard deviation requires that the

variable be interval. The variable "RS
HIGHEST DEGREE" is ordinal, which does
not meet this requirement.

The answer to the question is incorrect

application of a statistic.
 SW388R6

Problem 3
Data Analysis
and Computers I

Slide 16

Based on the dataset GSS2000.SAV, is the following statement

true, false, or an incorrect application of a statistic?

The range of 4 best represents the variability of the variable "RS

HIGHEST DEGREE".

1. True
2. False
3. Incorrect application of a statistic
 SW388R6

Solution 3
Data Analysis
and Computers I

Slide 17

The range requires that the

variable be ordinal or
interval. The variable "RS
HIGHEST DEGREE" is
ordinal, satisfying this
requirement.

The range best represents variability for

an ordinal level variable. The range for
the variable "RS HIGHEST DEGREE" is 4.

The answer to the question is true.

 SW388R6

Problem 4
Data Analysis
and Computers I

Slide 18

Based on the dataset GSS2000.SAV, is the following statement

true, false, or an incorrect application of a statistic?

The range for the variable "AGE OF RESPONDENT" is 16.81.

1. True
2. False
3. Incorrect application of a statistic
 SW388R6

Solution 4
Data Analysis
and Computers I

Slide 19

The range requires that the

variable be ordinal or
interval. The variable "AGE
OF RESPONDENT" is
interval, satisfying this
requirement.

The range for the variable "AGE OF

RESPONDENT" is 70, not 16.81.

The answer to the question is false.

 SW388R6

Problem 5
Data Analysis
and Computers I

Slide 20

Based on the dataset GSS2000.SAV, is the following statement

true, false, or an incorrect application of a statistic?

The variance for the variable "AGE OF RESPONDENT" is 70.

1. True
2. False
3. Incorrect application of a statistic
 SW388R6

Solution 5
Data Analysis
and Computers I

Slide 21

The variance requires that

the variable be interval. The
variable "AGE OF
RESPONDENT" is interval,
satisfying this requirement.

The variance for the variable "AGE OF

RESPONDENT" is 282.46, not 70.

The answer to the question is false.

 SW388R6

Problem 6
Data Analysis
and Computers I

Slide 22

Based on the dataset GSS2000.SAV, is the following statement

true, false, or an incorrect application of a statistic?

The standard deviation of 16.81 best represents the variability

of the variable "AGE OF RESPONDENT".

1. True
2. False
3. Incorrect application of a statistic
 SW388R6

Solution 6
Data Analysis
and Computers I

Slide 23

The standard deviation for

the variable "AGE OF
RESPONDENT" is 16.81.

The standard deviation The answer to the question

requires that the variable is true.
be interval. The variable
"AGE OF RESPONDENT" is
interval, satisfying this
requirement.

The standard deviation best represents variability for an interval

level variable whose distribution is not badly skewed. The
skewness for this variable is 0.595, not badly skewed by the
rule of thumb that the distribution of a variable is not badly
skewed if its skewness is between -1.0 and +1.0.
 SW388R6

Problem 7
Data Analysis
and Computers I

Slide 24

Based on the dataset GSS2000.SAV, is the following statement

true, false, or an incorrect application of a statistic?

The range of 22 best represents the dispersion of the variable

"NUMBER OF BROTHERS AND SISTERS".

1. True
2. False
3. Incorrect application of a statistic
 SW388R6

Solution 7
Data Analysis
and Computers I

Slide 25

The range requires that the

variable be ordinal or
interval. The variable
"NUMBER OF BROTHERS
AND SISTERS" is interval,
satisfying this requirement.
The range best represents
variability for an interval level
variable whose distribution is
badly skewed.

The skewness for this variable

is 2.391, badly skewed by the
rule of thumb that the
distribution of a variable is
badly skewed if its skewness
is less than or equal -1.0 or
greater than or equal to +1.0.

The range for the variable "NUMBER OF

BROTHERS AND SISTERS" is 22.

The answer to the question is true.

 SW388R6

Steps in solving variability problems - 1

Data Analysis
and Computers I

Slide 26

The following is a guide to the decision process for answering

central tendency homework problems.

If the problem asks about a specific measure of variability

Is the level of measurement No Incorrect application

required for the statistic of a statistic
satisfied?

Yes

Is the correct value for the No

statistic reported? False

Yes

True
 SW388R6

Steps in solving variability problems - 2

Data Analysis
and Computers I

Slide 27

If the problem asks about the preferred measure of variability:

Is the Is the
Is the No No
variable variable
variable
interval nominal
ordinal level?
level? level?

Yes Yes Yes

Is the value Is the

Yes SPSS does
for the range variable
Falsebadly not calculate
correct? the IQV
skewed?
No
Yes No

False True
Is the value for the
standard deviation or
variance correct?
No

Yes

False
True