Spurious Relationship - PowerPoint PPT Presentation

1 / 25
About This Presentation
Title:

Spurious Relationship

Description:

Spurious Relationship A correlation between two variables that have no causal relationship; both variables are caused by another variable Example: Water causes ... – PowerPoint PPT presentation

Number of Views:62
Avg rating:3.0/5.0
Slides: 26
Provided by: sobekCol
Category:

less

Transcript and Presenter's Notes

Title: Spurious Relationship


1
Spurious Relationship
  • A correlation between two variables that have no
    causal relationship both variables are caused
    by another variable

2
Example Water causes drunkenness.
  • Once a group of students decided to study
    empirically the causes of drunkenness.
  • They drank vodka and water. Got drunk.
  • They drank rum and water. Got drunk.
  • They drank scotch and water. Got drunk.
  • They drank bourbon and water. Got drunk.
  • They drank gin and water. Got drunk.
  • Conclusion Water causes drunkenness

3
Examples of spurious relationships
  • Amount of ice cream sold and death by drowning
    (units are months)
  • Number of doctors and number of people dying of
    disease (by city)
  • Number of churches and number of homicides (by
    city)
  • Number of libraries and number of people on drugs
    (over time)

4
What causes the price of rum?
  • There is a positive relationship between the
    salaries of Presbyterian ministers in
    Massachusetts and the price of rum in Havana.
  • What is the causal mechanism?

5
Causal Mechanism?
International business conditions prices and
salaries increasing everywhere
Ministers salaries
Price of rum
6
Popular Study Smoking causes bad grades
  • True or spurious?
  • What could be the variable (or variables) that
    drive both?

7
News reports
  • Bottled water causes healthier babies
  • Drinking tea makes you less likely to contract
    lung cancer

8
Research question Does education cause political
tolerance?
  • Many scholars have found a positive relationship
    between political tolerance (support for liberty
    for political enemies) and education.
  • Causal mechanism?
  • As people are educated, they are introduced to
    different kinds of people and perceived danger of
    political enemies decreases.
  • As people are educated, they are socialized to
    the correctness of constitutional rights.

9
Alternative Explanation
  • Political tolerance is actually caused by
    cognitive sophistication, as measured by IQ.
  • Causal mechanism?
  • It is difficult to reconcile the emotional
    reaction to named political enemies (i.e. Nazis)
    with constitutional rights and it takes a high IQ
    to understand that democracy benefits from giving
    liberties to bad people.

10
Relationship between Education and Political
Tolerance
High
Political Tolerance
Medium
Low
Low
Medium
High
Education
11
Relationship between Education and Political
Tolerance
High
Political Tolerance
Medium
Low
Low
Medium
High
Education
12
Controlling for IQ
  • X high IQ
  • Y low IQ

13
Education and Political Tolerance
High
Political Tolerance
Medium
Low
Low
Medium
High
Education
14
Education and Political Tolerance - High IQ
High
Political Tolerance
Medium
Low
Low
Medium
High
Education
15
Education and Political Tolerance - High IQ
High
Political Tolerance
Medium
Low
Low
Medium
High
Education
16
Education and Political Tolerance - Low IQ
High
Political Tolerance
Medium
Low
Low
Medium
High
Education
17
Education and Political Tolerance - Low IQ
High
Political Tolerance
Medium
Low
Low
Medium
High
Education
18
Conclusion?
  • What was the hypothesis?
  • What was the alternative explanation?
  • What do we conclude about the likely cause of
    political tolerance?

19
Review
  • A spurious effect is a specific case when
    correlation is not causation.
  • The specific case is related to whether the two
    things correlated are caused by a third variable.

20
An illustration with a causal model
A and B are spuriously correlated if
Z
B
A
21
A test of whether A and B are spurious
  • Look at the relationship between A and B,
    controlling for Z
  • This means, that for all values of Z, you should
    look at the relationship between A and B.

22
Crosstab analysis testing for spurious
relationships, part I (z 1 or 2)
23
Crosstab analysis testing for spurious
relationships, part 2 (z 1)
24
Crosstab analysis testing for spurious
relationships, part 3 (z 2)
25
Scatterplot illustration of a spurious
relationship
Z 1 or 2
A
Z 1
B
Z 2
A
A
B
B
Write a Comment
User Comments (0)
About PowerShow.com