This document discusses the challenges of establishing cause and effect relationships in psychology research. It explains that correlational research can identify relationships between variables but cannot prove causation, as correlations may be due to other unmeasured variables or reverse causation. Experimental research allows stronger claims about causation by manipulating an independent variable and measuring its effect on a dependent variable while controlling for other influences. However, not all variables can be experimentally manipulated. Overall, determining causes requires considering alternative explanations for observed relationships.
2. As a psychologist, when I hear people state a belief, I
immediately start thinking in terms of variables.
I realize that most people do not do this, but this is
ultimately the evidence we can collect and evaluate.
3. A variable is any quantity or quality that can take on
different values. Here are two random examples.
The Degree to which this Slideshow is Helpful.
Not at all Helpful Very Helpful
A Person’s Interest in Scientific Research
Not at all Interested Extremely Interested
4. Perhaps these two variables are related. Maybe
people who are more interested in the scientific
method will be more likely to find this slideshow to be
helpful.
5. Each person watching this slideshow can place themselves
along the continuum of each variable.
1. The degree to which this slideshow is helpful.
Not at all Helpful Very Helpful
2. A person’s interest in scientific research.
Not at all Interested Extremely Interested
6. Here is an example of one student, Phil.
Phil finds this to be a great slideshow. Phil
also loves science.
1. The degree to which this slideshow is helpful.
Phil
Not at all Helpful Very Helpful
2. A person’s interest in scientific research.
Phil
Not at all Interested Extremely Interested
7. Here is another example. This is where
James ranks himself on both variables.
1. The degree to which this slideshow is helpful.
James
Not at all Helpful Very Helpful
2. A person’s interest in scientific research.
James
Not at all Interested Extremely Interested
8. We could ask a large group of students to mark where
they fall on both variables.
Perhaps we might start to see a pattern. People who
like science keep ranking this slideshow as helpful.
People who do not like science rank it lower.
9. It is very tempting to start fitting these findings into a
story. Maybe people who like science are more
curious and patient. Maybe they find slideshows
(which are, admittedly, kind of dull) to be more
interesting.
11. How do we research Cause and Effect relationships in
Psychology?
12. Establishing Cause and Effect
John Stuart Mill David Hume Immanuel Kant
Philosophers have a lot to say about how we
determine if one thing causes another. We will skip
that debate and focus on three characteristics to
help us establish cause and effect.
13. Establishing Cause and Effect
Conditions to establish Cause and Effect.
1. The variables are correlated.
2. The cause comes before the effect.
3. There are no other variables to explain the effect.
14. Establishing Cause and Effect
Note that the first step is to establish a correlation.
Correlation is a fancy term for “related.” To understand
how psychologists study causes and effects, we
need to understand correlation.
15. Establishing Cause and Effect
A B
It is very easy to think about how one thing may cause
another. However, it is not always easy to see how
these ideas imply a correlation. Let’s look at
examples.
16. Correlation examples
Does having a nice teacher cause students to learn
more? Here are the two variables:
Teacher’s Niceness
Not at all Nice Very Nice
Student Learning
Very Little Learning Mastery of Content
17. Correlation examples
If the variables are related, then we should see nicer
teachers have classes that perform better. Pretend
we observed a teacher that fits this pattern, Mr.
Carter.
Teacher’s Niceness
Mr.
Carter
Not at all Nice Very Nice
Student Learning
Mr. Carter’s
Class
Very Little Learning Mastery of Content
18. Correlation examples
If the variables are related, then we also should see
teachers who are not nice have classes that perform
poorly. Pretend we observed a teacher like this, Ms.
Stark.
Teacher’s Niceness
Ms.
Stark
Not at all Nice Very Nice
Student Learning
Ms. Stark’s
Class
Very Little Learning Mastery of Content
19. Two Sides to a Correlation
This is the hard part. Correlations have two sides, or
ends. People usually only think about one. In our
case, we think about good teachers having good
students (the high end).
20. Two Sides to a Correlation
If you believe a nice teacher causes a class to do
better then you are also saying that not being nice
goes with students doing worse. This is the low end
of the correlation. For a correlation to exist, we need
to have both Mr. Carter and Ms. Stark!
21. More Examples – Two Sides of Correlation
If you believe that giving someone flowers will make
them like you, then you also should observe that
people to whom you do not give flowers would like
you less.
22. More Examples – Two Sides of Correlation
If you believe that spanking a child causes the child to
behave better, then you are also saying that not
spanking should co-occur with worse behavior.
23. Each of these is a proposed correlation. It can be
difficult to think about these, but practicing this can
help you think about what type of observations you
would need in order to see if two variables are
correlated.
24. Correlational Methods
Most correlational research methods involve
collecting data on at least two variables.
Surveys /
There are many ways to collect these data. Questionnaires
See your text for those details.
Archival Research
Naturalistic Observation
26. Establishing Cause and Effect
The problems are the second and third conditions
Conditions to Establish Cause and Effect
1. The Variables are Correlated.
2. The Cause comes before the Effect.
3. There are no other variables to explain the effect.
27. Problem
2. The Cause comes before the Effect
Usually, in correlational studies, we observe both
variables at the same time. It is not always possible
to know which comes first.
28. Problem
2. The Cause comes before the Effect
For the example with Mr. Carter and Ms. Stark, this is
a problem. The idea that a nice teacher causes a
class to learn more is only one possible direction of
cause between these variables. As you see below,
this explanation hypothesizes that the niceness
comes first.
Teacher’s Causes Students’
Niceness Learning
29. Problem
2. The Cause comes before the Effect
The opposite direction is also possible. Some
teachers may respond to their class’s performance.
It feels great to have a class do well, and it can feel
very defeating if your class is struggling and
performing poorly.
Teacher’s Causes Students’
Niceness Learning
30. Problem
2. The Cause comes before the Effect
For every correlation you observe, be sure to consider
both possible directions of cause.
Variable A Causes Variable B
OR
Variable A Causes Variable B
32. Establishing Cause and Effect
The third condition also has to be met.
Conditions to Establish Cause and Effect
1. The Variables are Correlated.
2. The Cause comes before the Effect.
3. There are no other variables to explain the effect.
33. Problem
3. Are there other variables to explain the effect?
In correlational research, even if we observe a
correlation and establish that one variable happens
before the other one…
…we still do not know if there is a cause and effect
relationship.
34. Problem
3. Are there other variables to explain the effect?
The problem is that correlational research observes
variables as they occur in the real world. Although
valuable, this leaves many questions.
35. Problem
3. Are there other variables to explain the effect?
In the real world, there are countless other variables.
Some may also be recorded in our study, but others
may vary without us knowing.
36. Problem
3. Are there other variables to explain the effect?
In the most simple form, this is called the third-variable
problem. Does some other variable, other than our
two observed variables, cause the other two to be
related?
Third
Variable
Variable A Variable B
37. Problem
3. Are there other variables to explain the effect?
For our teaching example, pretend that we found a
correlation between teacher’s niceness and student
learning (step 1 in our search for causes).
Teacher’s Student
Niceness Learning
38. Problem
3. Are there other variables to explain the effect?
Next, pretend that we establish that the teacher’s
niceness comes first (step 2 in our search for
causes).
39. Problem
3. Are there other variables to explain the effect?
We still would not be able to say for certain that
another variable did not cause the observed
correlation.
Third
Variable ?
Teacher’s Student
Niceness Learning
40. Problem
3. Are there other variables to explain the effect?
For example, maybe the resources available to the
school affects both of the other variables.
School
Resources
Teacher’s Student
Niceness Learning
41. Problem
3. Are there other variables to explain the effect?
Let’s look at another example. There is a correlation
between the number of churches in a city and the
amount of crime. Cities with more churches have
more crime. Yes, that’s right, more churches is
associated with more crime!
Number of Number of
Churches in Violent Crimes
City in City
42. Don’t worry, this correlation is meaningless.
It is explained by a third variable.
43. Problem
3. Are there other variables to explain the effect?
There is a simple third variable: Population of the City.
The more people in a city, the more churches they
build. The more people in a city, the more chances
there are for crime.
Population
Size of City
Number of Number of
Churches in Violent Crimes
City in City
44. Review
Correlational research is a valuable tool.
There are many ways to collect correlational data.
45. Review
Correlation is a first step in the search for causes.
To establish cause, we need all three of the following
conditions:
1. The Variables are Correlated.
2. The Cause comes before the Effect.
3. There are no other variables to explain the
effect.
46. Review
Simply observing a correlation does not tell us that
one variable causes the other.
The direction of cause could go either way.
There may be a third variable causing the observed
variables to look correlated.
48. The Experimental Method
The experimental method is a researcher’s strongest
tool for establishing cause and effect relationships.
A well-structured experiment meets all three criteria
for establishing cause and effect.
49. The Experimental Method
Conditions to Establish Cause and Effect
1. The Variables are Correlated.
2. The Cause comes before the Effect.
3. There are no other variables to explain the effect.
50. The Experimental Method
1. The experimenter looks for a relationship between
one variable that is a cause and another variable
that shows the effect.
2. The experimenter manipulates the causal variable
first and measures the effect variable later.
3. By using manipulation and controls, the
experimenter can rule out alternate explanations.
51. The Experimental Method
The two variables have
special names.
Independent variable – the
variable manipulated by
the researcher to see if it
is a cause.
Dependent variable – the
variable measured to see
if there is an effect.
52. The Experimental Method
Independent Variables
Typically, independent variables describe what is
different between two groups:
an experimental group and control group.
53. The Experimental Method
Independent Variables
Perhaps the hardest part of conducting an experiment
is making sure that the independent variable, and
ONLY the independent variable, differs between the
groups.
54. The Experimental Method
Independent Variables, Example
We could use an experiment to study the effect of
teacher niceness on student learning. We would
need to create two groups of students:
55. The Experimental Method
Independent Variables, Example
Experimental Group – Students assigned
a teacher who has instructions in
specific ways to be nice in class.
Control Group – Students assigned a
teacher without instructions to be
nice.
56. The Experimental Method
Independent Variables, Example
The hard part is making sure that the only thing that
differs systematically between these two groups is
the niceness of the teacher.
57. The Experimental Method
Ways to Eliminate Extraneous Variables
An extraneous variable is a variable other than the
independent variable that might be affecting the
dependent variable.
58. The Experimental Method
Ways to Eliminate Extraneous Variables
There are two main ways to eliminate Extraneous
Variables:
1. Experimental Control – designing the study
carefully to remove extraneous variables.
2. Randomization – If an extraneous variable cannot
be removed, it can be randomly distributed across
experimental groups. That way it does not vary
systematically with the independent variable.
59. Summary
Give yourself ample time to master these ideas and
the terminology. You should be able to define all of
the following terms, and describe how they relate to
each other.
Experiment Extraneous Variable
Independent Variable Experimental Control
Dependent Variable Randomization
Experimental Group
Control Group
60. Limits of the Experimental Method
Not every variable can be manipulated. Remember
the first example of two correlated variables:
The Degree to which this Slideshow is Helpful.
Not at all Helpful Very Helpful
A Person’s Interest in Scientific Research
Not at all Interested Extremely Interested
61. Limits of the Experimental Method
We hypothesized that a person’s interest in science
will affect how useful they think this slideshow is.
Interest in
Causes Perception of
Scientific
this Slideshow
Research
62. Limits of the Experimental Method
How would you manipulate people’s “interest in
science” for an experiment, making some people
very interested and others disinterested?
? ?
Interest in
Scientific
Research
63. Limits of the Experimental Method
Even if we could come up with a creative way to
manipulate “interest in science,” would it be the
same thing as the interest people develop over a
lifetime?
64. Limits of the Experimental Method
Other variables cannot be manipulated without
causing harm or other problems.
65. Limits of the Experimental Method
Summary
Some variables are impractical to manipulate.
Experiments can be artificial.
Some experiments cannot be conducted ethically.
66. Review
Practice turning explanations or stories into
hypotheses about the relationship between two
variables.
When variables are correlated, think about both
ends, or poles of the relationship: The high end, and
the low end.
67. Review
Conditions to establish Cause and Effect.
1. The variables are correlated.
2. The cause comes before the effect.
3. There are no other variables to explain the effect.
68. Review
Correlational Research Methods can identify
relations between variables that you measure.
Correlational Research cannot offer strong evidence
of cause and effect!
69. Review
Correlations have several possible causal
explanations
1. The direction of cause could go either way.
2. A third variable could cause both variables.
70. Review
The Experimental Method allows us to establish
cause and effect relationships by manipulating an
independent variable in a controlled setting, then
looking for its effect on a dependent variable.
71. These are the foundational ideas to help you think
about research more critically.
There are many more details in your book, which can
help to improve your ability to evaluate evidence
about causes and effects.