An Introduction to the Use and Misuse of Research Variables

1

• An Introduction to the Use and Misuse of Research
  Variables

A judicious man uses statistics, not to get
knowledge, but to save himself from having
ignorance foisted upon him. --Thomas Carlyle
2
Variable Defined
Any entity that can take on different values.
Age is a variable because it can take different
values for different people at different times.
3
Examples

• Blood Pressure
• Sex
• Gender
• Age
• Extraversion
• Patient Satisfaction
• Heart rate
• Political Party

• Time
• Weight
• Height
• Anxiety
• Pleasure
• Fear
• Aggression
• Attractiveness

4
The Aspects of Variables
5
Variables
Attributes
Values
Relationship
6
1. Attributes
The specific value of a variable.
7
Examples of Attributes

• SEX
• Male
• Female

• Gender
• Assertive
• Responsive
• Androgynous

• Math makes me anxious
• Agree
• Neutral
• Disagree

8
Values
The numerical aspect directly associated with a
specific attribute.
9
Examples of Values

• SEX
• Male 1
• Female 2

• Gender
• Assertive 81lt 120
• Responsive 0 lt 40
• Androgynous 41lt80

• Math makes me anxious
• Agree 1
• Neutral 2
• Disagree 3

10
Relationship
Does a statistical relationship exist between the
attributes of the variable?
The correspondence between two variables.
11
Types of Relationships
12
For Example
You do not become more female if your male level
goes down. They are independent of each other.
However, your level of assertiveness goes down as
your responsiveness level goes up (negative
relationship).
13
Levels of Measurement
14
Nominal Measures
Attributes are only named. Also, referred to as
categorical measures.
15
For Example

• SEX
• Male
• Female

• Ethnicity
• Caucasian
• African American
• Middle Eastern
• Indian
• Native American
• Other

• Had a heart attack?
• Yes
• No

16
3 Characteristics of Nominal Variables
1. Must be mutually exclusive.
2. Must be equivalent. (no comparing apples to
oranges)
3. Must be exhaustive.
17
What is the problem with this?
18
Republicans, Democrats, and Independents are not
exhaustive of all possible political parties.
How would you correct this?
The simplest way would be to add an other
category.
19
Ordinal Measures
Attributes can be ordered.
20
For Example

• SES
• Lower Class
• Middle Class
• Upper Class

• Education Level
• Grade School
• Middle School
• High School
• Junior College
• College/University Graduate
• Post Graduate

• Letter Grade
• A
• B
• C
• D
• F

21
BIG NOTE
We do not know the magnitude of the difference
between the variables. We just know there is a
clear difference.
22
In Other Words
If Middle Class is 30-50,000 And Upper Class is
50,000
Then The difference between the two could be
pennies or millions of dollars We dont know.
23
Interval Measures
The distance between attributes has real
quantitative meaning.
24
The distance between IQs of 80 100 is the same
as the distance between IQs of 140 160.
25
However, we cannot say that someone with an IQ of
160 is twice as intelligent as someone with an IQ
of 80.
This would require an Absolute Zero.
26
General Examples

• Patient Compliance
• Temperature (Celsius and Fahrenheit)
• Neuroticism
• Extraversion
• Patient Satisfaction
• Physician Humor Orientation

27
Scaled Interval Measures
- Likert
- Semantic Differential
- Thurstone
28
Likert Scales
5 Strongly Agree 4 Agree 3 Undecided 2
Disagree 1 Strongly Disagree
_____1. My physician regularly communicates with
others joking with them. _____2. People usually
do not laugh at my physicians attempts at humor.
29
Semantic Differential Scales
Numbers 1 and 7 indicate a very strong feeling.
Numbers 2 and 6 indicate a strong feeling.
Numbers 3 and 5 indicate a fairly weak feeling.
Number 4 indicates you are undecided.
30
Thurstone Scales
31
Important Scale Characteristics
1. Multiple questions measuring one concept.
2. Scales must be statistically reliable and
valid.
3. Associated numerical values must be
meaningful.
32
Ratio Measures
The distance between attributes has real
quantitative meaning, and has an absolute zero.
33
Celsius and Fahrenheit are Interval measures for
temperature.
Kelvin is a ratio measures for temperature
because it has an absolute zero.
34
Examples

• Temperature (Kelvin)
• Age
• Height
• Weight
• Mass
• Blood Pressure
• Speed
• Heart Rate

35
(No Transcript)
36
Why Do Levels of Measurement Matter?
The type of variables you have determine the
statistical devices you can and cannot use.
37
See Picking a Test
38
Variables in Research
39
Dependent Variables
The variable that is effected or not effected by
another variable in a research study.
40
Independent Variables
The variable that is being manipulated or
examined in a study to see if it effects the
dependent variable.
41
Example
In a study by Wrench and Booth-Butterfield
(2001), it was found that a physicians humor
orientation influenced her or his patients level
of satisfaction.
42
What was the IV and DV?
IV Physician Humor Orientation
DV Patient Satisfaction
43
Another Example
Wrench and Booth-Butterfield (2001) also found
that the severity of a diagnosis influenced a
patients desire for information about the
diagnosis and information about treatment, but
did not effect patient compliance.
44
What were the IVs and DVs?

• IV Severity
• DV Desired Information about
  Diagnosis

2) IV Severity DV Desired Information
about Treatment
45
What were the IVs and DVs? Cont
3) IV Severity DV Patient Compliance
46
Last Example
In a study by Wrench (2002), it was found that
males were more physically and verbally
aggressive than females.
47
What were the IVs and DVs?

• IV Sex (Males Females)
• DV Physical Aggression

2) IV Sex (Males Females) DV Verbal
Aggression