TA: Clint Inkster

1
TA Clint Inkster office hour Thursday
1130-1230 N2012 email inks1715_at_wlu.ca
2
Basic Research Designs
The Operational Definition
Many phenomena of interest to psychology are
internal mental states
In order to be objective, we need to identify
external, observable events that are valid and
reliable indicators that the internal event is
taking place
3
Basic Research Designs
The Operational Definition
How would you operationalize the following
phenomena?
a) hunger
b) depression
c) sexual attraction
d) intelligence
4
Basic Research Designs
1. Descriptive Research
- surveys, interviews, case studies
- naturalistic observation
5
Basic Research Designs
2. Correlational Designs
- Pearson Product-Moment Correlation (r)
- evaluates the strength of relationship between
2 observed variables
6
Basic Research Designs
2. Correlational Designs
- Pearson Product-Moment Correlation (r)
- evaluates the strength of relationship between
2 observed variables
- no info regarding cause and effect
7
Correlation
height
variable x
weight
variable y
8
Basic Research Designs
2. Correlational Designs
- there is a positive correlation between height
and weight
- as the value of one variable increases the
value of the other tends to increase
- relationship is not a perfect correlation
- correlations can be positive or negative
- zero correlation means no relationship
9
Basic Research Designs
2. Correlational Designs
correlation and cause and effect
- we have identified a strong positive
relationship between height and weight
- does being taller cause you to be heavier?
- does being heavier cause you to be taller?
10
Basic Research Designs
2. Correlational Designs
correlation and cause and effect
- correlational evidence gives us no basis to
choose between causal explanations
- in this case both are probably wrong
- 3rd variable problem (e.g., nutrition)
11
Basic Research Designs
2. Correlational Designs
correlation and cause and effect
Does smoking cause lung cancer?
- this claim can probably be made for rats,
monkeys, etc...
- no causal evidence that smoking causes lung
cancer in humans (although there certainly
appears to be a strong relationship)
12
Basic Research Designs
3. Experimental Designs
- t-tests, ANOVA, chi-square, etc...
13
Basic Research Designs
3. Experimental Designs
- t-tests, ANOVA, chi-square, etc...
14
Basic Research Designs
3. Experimental Designs
Variables
a variable is anything that is free to take on
different values (if it never changes, it is a
constant, not a variable)
Dependent variable (DV) what gets measured
Independent variable (IV) what gets manipulated
15
Basic Research Designs
3. Experimental Designs
Variables
In the simplest case, there is one DV and one IV
Ideally, all other potential variables are held
constant (experimental control)
If our control procedures are effective, the only
possible explanation for changes in the DV is our
manipulation of the IV
16
Basic Research Designs
3. Experimental Designs
Variables
Even when there is a single IV, an experiment may
involve several levels of that IV
e.g., a study of drug tolerance may include
several groups, each getting a different dose of
the drug
17
Basic Research Designs
3. Experimental Designs
Variables
IV drug dose
levels of IV
0 mg/kg (control group)
1 mg/kg
10 mg/kg
100 mg/kg
18
Basic Research Designs
3. Experimental Designs
Control
Hypothesis students will learn better
(operationalized as higher grades) in brightly
lit classrooms than they will in darkened
classrooms
DV
grades
IV
classroom illumination
19
Basic Research Designs
3. Experimental Designs
Control
Hypothesis students will learn better
(operationalized as higher grades) in brightly
lit classrooms than they will in darkened
classrooms
Group A PS291 in brightly lit room
Group B PS291 in dimly lit room
20
Basic Research Designs
3. Experimental Designs
Control
Hypothesis students will learn better, as
evidenced by higher grades, in brightly lit
classrooms than they will in darkened classrooms
What potential confounding variables need to be
controlled for in this study?
21
Basic Research Designs
3. Experimental Designs
Control
Where possible, we hold all extraneous variables
constant (e.g., both classes at the same time of
day)
Where we cant control the variable directly, we
rely on randomization
- random assignment of subjects to groups where
possible
22
Basic Research Designs
3. Experimental Designs
Validity
Internal validity
Does our experiment really measure what we think
it is measuring?
- does the WAIS measure intelligence?
- does hypnotic age regression work?
- does the shape of your head indicate criminal
tendencies?
23
Basic Research Designs
3. Experimental Designs
Validity
External validity
Does what we see in the laboratory apply out
there in the real world?
- would every member of the population respond to
a given manipulation in the same way as the
samples in my study?
- generalizability
24
Basic Research Designs
Inductive and Deductive Reasoning
Inductive reasoning involves starting from
specific facts and leading to general theories
about a class of events
e.g., specific observations of the behavior of
newborn ducklings leads to a general theory of
imprinting
25
Basic Research Designs
Inductive and Deductive Reasoning
Deductive reasoning involves starting from a
general theory and generating specific
predictions about a specific situation
e.g., theory of imprinting leads to prediction
that newborn ducklings will imprint on a chicken
if it is the first creature they see after
hatching
26
Basic Research Designs
The Arch of Knowledge
theory
deduction
induction
facts / observations
27
Basic Research Designs
More Propositional Logic
Modus Ponens (confirmatory)
P1 if p, then q
P2 p
C q
This is a logically valid form of argument
28
Basic Research Designs
More Propositional Logic
Modus Ponens (confirmatory)
P1 if it rains, then the ground will be wet
P2 it is raining
C the ground will be wet
This is a logically valid argument
29
Basic Research Designs
More Propositional Logic
Affirming the Consequent
P1 if p, then q
P2 q
C p
This is not a logically valid argument
30
Basic Research Designs
More Propositional Logic
Affirming the Consequent
P1 if it rains, then the ground will be wet
P2 the ground is wet
C it is raining
This is not a logically valid argument
31
Basic Research Designs
More Propositional Logic
Modus Tollens (disconfirmatory)
P1 if p, then q
P2 not q
C not p
This is a logically valid form of argument
32
Basic Research Designs
More Propositional Logic
Modus Tollens (disconfirmatory)
P1 if it rains, then the ground will be wet
P2 the ground is not wet
C it is not raining
This is a logically valid argument
33
Basic Research Designs
More Propositional Logic
Denying the Antecedent
P1 if p, then q
P2 not p
C not q
This is not a logically valid argument
34
Basic Research Designs
More Propositional Logic
Denying the Antecedent
P1 if it rains, then the ground will be wet
P2 it is not raining
C the ground is not wet
This is not a logically valid argument
35
Basic Research Designs
A Basic Logic Problem
- each card has a circle (with or without an x)
on one side and a triangle (with or without an
x) on the other side
- which 2 cards would you have to turn over to
test the following hypothesis
hypothesis if a card has a circle with an x
on one side, it must have a triangle without an
x on the other side
36
Basic Research Designs
A Basic Logic Problem
Answer (part 1) Obviously we need to flip Card A
- according to our hypothesis, Card A must have a
triangle without an x on the other side
- this is a test to confirm the hypothesis (i.e.,
modus ponens)
hypothesis if a card has a circle with an x
on one side, it must have a triangle without an
x on the other side
37
Basic Research Designs
A Basic Logic Problem
Answer (part 2) Not so obviously, we also need
to flip Card B
- according to our hypothesis, Card B cannot have
a circle with an x on the other side
- this is a test to disconfirm the hypothesis
(i.e., modus tollens)
hypothesis if a card has a circle with an x
on one side, it must have a triangle without an
x on the other side
38
Basic Research Designs
A Basic Logic Problem
Most people faced with this problem pick A and D
- there is nothing in the rules or our
hypothesis that says that only cards with a
circle/x on one side can have a triangle/no x on
the other side
- this is denying the antecedent (invalid logic)
hypothesis if a card has a circle with an x
on one side, it must have a triangle without an
x on the other side
39
Next Lecture
Chapter 3 (all) Developing the
Hypothesis Chapter 4 (pp 81-86) Scales of
Measurement