convergent validity

Chapter 2
Personality Research Methods
Measuring personality variables
 What is a variable?
 What kinds of variables are there?
– Categorical variables (examples: gender, political affiliation)
– Continuous variables (height, weight)
 How can we measure an aspect of personality as a
variable?
 Psychometrics (psyche – mind, metric – measure): an
applied branch of psychology that deals with psychological measurement
 The most commonly used approach to study human
personality: the self-report questionnaire measure
Personality Inventory
Please read the following items and decide how well they describe your
personality. Respond to each according to the following list of alternatives:
a.
b.
c.
d.
1.
2.
3.
4.
5.
6.
7.
8.
9.
very much unlike me
somewhat unlike me
somewhat like me
very much like me
I am socially somewhat awkward.
I don’t find it hard to talk with strangers
I feel tense when I’m with people I don’t know well.
When conversing I worry about saying something dumb.
I feel nervous when speaking to someone in authority.
I am often uncomfortable at parties and other social functions.
I feel inhibited in social situations.
I have trouble looking someone right in the eye.
I am more shy with members of the opposite sex.
Correlating variables
 Correlating two variables
– Measure the two variables (X and Y) as accurately as
necessary
– Plot the variables as X-Y coordinates on a twodimensional space
– Examine the shape of the “scatterplot” to estimate the
size and “direction” of the resulting correlation
– Compute the correlation coefficient using the correct
mathematical formula (or let the computer do it for
you)
NAME
HEIGHT (X)
WEIGHT (Y)
Allport, Allen
69
175
Bimiup, Scotty
68
159
Brewer, Joseph
71
177
Derrickson, Derek
68
150
Dorman, Carlton
67
162
Horne, Karyn
65
134
Ingram, Irene
64
121
James, Jocelyn
67
125
Jones, Ernie
67
158
Lang, Fred
69
161
Kelly, Christine
63
119
Lime, Nora
68
128
etc.
Reliability and validity in personality
measurement
 Reliability (consistency of measurement)
– Internal consistency
• Item-to-total correlation
• Split-half reliability
• Cronbach’s alpha
– Test-retest reliability
 Validity (construct validity)
– Face validity
– Concurrent validity
– Convergent and discriminant validity
– Criterion (predictive) validity
Personality Inventory
Please read the following items and decide how well they describe your
personality. Respond to each according to the following list of alternatives:
a.
b.
c.
d.
1.
2.
3.
4.
5.
6.
7.
8.
9.
very much unlike me
somewhat unlike me
somewhat like me
very much like me
I am socially somewhat awkward.
I don’t find it hard to talk with strangers
I feel tense when I’m with people I don’t know well.
When conversing I worry about saying something dumb.
I feel nervous when speaking to someone in authority.
I am often uncomfortable at parties and other social functions.
I feel inhibited in social situations.
I have trouble looking someone right in the eye.
I am more shy with members of the opposite sex.
Reliability and validity in personality
measurement
 Reliability (consistency of measurement)
– Internal consistency
• Item-to-total correlation
• Split-half reliability
• Cronbach’s alpha
– Test-retest reliability
 Validity (construct validity)
– Face validity
– Concurrent validity
– Convergent and discriminant validity
– Criterion (predictive) validity
Item-to-total correlations
SHY1 SHY2 SHY3 SHY4 SHY5 SHY6 SHY7 SHY8 SHY9
SHY .68
.64
.67
.67
.65
.73
.75
.57
.61
Reliability and validity in personality
measurement
 Reliability (consistency of measurement)
– Internal consistency
• Item-to-total correlation
• Split-half reliability
• Cronbach’s alpha
– Test-retest reliability
 Validity (construct validity)
– Face validity
– Concurrent validity
– Convergent and discriminant validity
– Criterion (predictive) validity
Reliability and validity in personality
measurement
 Reliability (consistency of measurement)
– Internal consistency
• Item-to-total correlation
• Split-half reliability
• Cronbach’s alpha
– Test-retest reliability
 Validity (construct validity)
– Face validity
– Concurrent validity
– Convergent and discriminant validity
– Criterion (predictive) validity
Reliability (internal consistency)
Cronbach’s alpha = .83
Factor pattern
Factor 1
SHY1
.69
SHY2
.62
SHY3
.67
SHY4
.66
SHY5
.65
SHY6
.75
SHY7
.78
SHY8
.54
SHY9
.58
Reliability and validity in personality
measurement
 Reliability (consistency of measurement)
– Internal consistency
• Item-to-total correlation
• Split-half reliability
• Cronbach’s alpha
– Test-retest reliability
 Validity (construct validity)
– Face validity
– Concurrent validity
– Convergent and discriminant validity
– Criterion (predictive) validity
Reliability and validity in personality
measurement
 Reliability (consistency of measurement)
– Internal consistency
• Item-to-total correlation
• Split-half reliability
• Cronbach’s alpha
– Test-retest reliability
 Validity (construct validity)
– Face validity
– Concurrent validity
– Convergent and discriminant validity
– Criterion (predictive) validity
Personality Inventory
Please read the following items and decide how well they describe your
personality. Respond to each according to the following list of alternatives:
a.
b.
c.
d.
1.
2.
3.
4.
5.
6.
7.
8.
9.
very much unlike me
somewhat unlike me
somewhat like me
very much like me
I am socially somewhat awkward.
I don’t find it hard to talk with strangers
I feel tense when I’m with people I don’t know well.
When conversing I worry about saying something dumb.
I feel nervous when speaking to someone in authority.
I am often uncomfortable at parties and other social functions.
I feel inhibited in social situations.
I have trouble looking someone right in the eye.
I am more shy with members of the opposite sex.
Reliability and validity in personality
measurement
 Reliability (consistency of measurement)
– Internal consistency
• Item-to-total correlation
• Split-half reliability
• Cronbach’s alpha
– Test-retest reliability
 Validity (construct validity)
– Face validity
– Concurrent validity
– Convergent and discriminant validity
– Criterion (predictive) validity
Example of concurrent validity
Correlate the participants’ total shyness score for the 9-item measure of
shyness (X variable) with their rating on the following dimensional
scale (Y variable):
In general, how shy are you?
_________________________________________________________
0
10
20
30
40
50
60
70
80
90 100
If both X and Y are good alternative ways to measure shyness, the
correlation between X and Y should be positive and large (for
example, r = .86).
Reliability and validity in personality
measurement
 Reliability (consistency of measurement)
– Internal consistency
• Item-to-total correlation
• Split-half reliability
• Cronbach’s alpha
– Test-retest reliability
 Validity (construct validity)
– Face validity
– Concurrent validity
– Convergent and discriminant validity
– Criterion (predictive) validity
Example of convergent and discriminant
validity
Correlate the participants’ total shyness score for the 9-item measure of
shyness (X variable) with their total scores on the following other
personality measures:
Shyness score correlated with:
Social anxiety
.91
Introversion
.66
Conscientiousness
.15
IQ
.03
The first two correlations provide evidence of the shyness measure’s
convergent validity. The last two correlations provide evidence of the shyness
measure’s discriminant validity.
Reliability and validity in personality
measurement
 Reliability (consistency of measurement)
– Internal consistency
• Item-to-total correlation
• Split-half reliability
• Cronbach’s alpha
– Test-retest reliability
 Validity (construct validity)
– Face validity
– Concurrent validity
– Convergent and discriminant validity
– Criterion (predictive) validity
Four possible causal structures of an A-B
relationship
A
B
Changes in Variable A cause corresponding changes to occur
in Variable B
B
A
Changes in Variable B cause corresponding changes to occur
in Variable A
A
B
Changes in Variable A cause changes in Variable B, which
in turn cause changes in Variable A, which in turn . . .
A
C
B
Changes in Variable C cause corresponding changes in both
Variables A and B. Variables A and B therefore covary even
though neither one causes the other
Three categories of moderating variables in
personality research
Category
Function
Representative Example
Situational
Specify in which types of
moderating variables situations traits will be good
versus poor predictors of their
trait-relevant behaviors
Psychologically weak
versus strong situations
Personal moderating
variables
Private self-consciousness
Specify for which types of
people traits will be good
versus poor predictors of their
trait-relevant behaviors
Criterion moderating Specify the types of behaviors
variables
that traits will predict either
well or poorly
Level of aggregation of
the behavioral measure
Experimental study
Correlational study
Goal
Test for a causal relationship
between a manipulated variable
and a measured variable
Test for a correlational relationship
between two measured variables
Variables
Independent variable (manipulated)
Dependent variable (measured)
Predictor variable (measured)
Criterion variable (measured)
Criteria
Experimenter controls the IV
Random assignment to conditions
Measures of both the predictor and
the criterion variables must be
obtained
Major pitfalls
Confounded manipulation of the
independent variable
Confounded measurement of the
dependent variable
Using a measured variable that is
unreliable or that lacks validity
Confounded measurement of the
predictor variable
Confounded measurement of the
criterion variable
Using measured variables that are
unreliable or that lack validity
Possible
conclusions
Level of IV affects level of DV
Level of IV doesn’t affect level of
DV
PV is related to (associated with) CV
PV is not related to CV
Best safeguard
Replication
Replication
Methods for measuring aspects of personality
 Trait measures (e.g., self-report questionnaires)
 State experience measures (e.g., mood ratings)
 Ability tests (e.g., intelligence test)
 Subjective ratings of behavior (e.g., self-ratings of one’s
own behavior)
 Objective ratings of behavior (e.g., observer ratings of
one’s behavior)
 Behavioral measures (e.g., counts of specific behaviors)
 Physiological measures (e.g., galvanic skin response)
Personality Inventory
Please read the following items and decide how well they describe your
personality. Respond to each according to the following list of alternatives:
a.
b.
c.
d.
1.
2.
3.
4.
5.
6.
7.
8.
9.
very much unlike me
somewhat unlike me
somewhat like me
very much like me
I am socially somewhat awkward.
I don’t find it hard to talk with strangers
I feel tense when I’m with people I don’t know well.
When conversing I worry about saying something dumb.
I feel nervous when speaking to someone in authority.
I am often uncomfortable at parties and other social functions.
I feel inhibited in social situations.
I have trouble looking someone right in the eye.
I am more shy with members of the opposite sex.
Methods for measuring aspects of personality
 Trait inventories (e.g., self-report questionnaires)
 State experience measures (e.g., mood ratings)
 Ability tests (e.g., intelligence test)
 Subjective ratings (e.g., self-ratings of behavior)
 Objective ratings (e.g., observer ratings of behavior)
 Behavioral measures (e.g., counts of specific behaviors)
 Physiological measures (e.g., galvanic skin response)
Example of a Mood Adjective Check List
Check all of the following adjectives that describe your
current mood:
____ anxious
____ confident
____ agitated
____ confused
____ serene
____ nervous
____ apprehensive
____ relaxed
____ calm
OR . . . To what extent do you feel anxious?
0
not at all
1
2
3
slightly moderately extremely
Methods for measuring aspects of personality
 Trait inventories (e.g., self-report questionnaires)
 State experience measures (e.g., mood ratings)
 Ability tests (e.g., intelligence test)
 Subjective ratings (e.g., self-ratings of behavior)
 Objective ratings (e.g., observer ratings of behavior)
 Behavioral measures (e.g., counts of specific behaviors)
 Physiological measures (e.g., galvanic skin response)
IQ tests are cognitive ability tests
Methods for measuring aspects of personality
 Trait inventories (e.g., self-report questionnaires)
 State experience measures (e.g., mood ratings)
 Ability tests (e.g., intelligence test)
 Subjective ratings (e.g., self-ratings of behavior)
 Objective ratings (e.g., observer ratings of behavior)
 Behavioral measures (e.g., counts of specific behaviors)
 Physiological measures (e.g., galvanic skin response)
Subjective self-ratings of behaviors
1. To what extent did you take the lead during the
conversation?
not at all
somewhat
moderately
very much
2. To what extent did you disclose personal information
during the conversation?
not at all
somewhat
moderately
very much
3. To what extent did you appear self-conscious during the
conversation?
not at all
somewhat
moderately
very much
Observing and rating people’s behavior
Objective observer ratings of behaviors
1. To what extent did Participant A take the lead during the
conversation?
not at all
somewhat
moderately
very much
2. To what extent did Participant A disclose personal
information during the conversation?
not at all
somewhat
moderately
very much
3. To what extent did Participant A appear self-conscious
during the conversation?
not at all
somewhat
moderately
very much
Methods for measuring aspects of personality
 Trait inventories (e.g., self-report questionnaires)
 State experience measures (e.g., mood ratings)
 Ability tests (e.g., intelligence test)
 Subjective ratings (e.g., self-ratings of behavior)
 Objective ratings (e.g., observer ratings of behavior)
 Behavioral measures (e.g., counts of specific behaviors)
 Physiological measures (e.g., galvanic skin response)
Observing and recording behavioral measures
Methods for measuring aspects of personality
 Trait inventories (e.g., self-report questionnaires)
 State experience measures (e.g., mood ratings)
 Ability tests (e.g., intelligence test)
 Subjective ratings (e.g., self-ratings of behavior)
 Objective ratings (e.g., observer ratings of behavior)
 Behavioral measures (e.g., counts of specific behaviors)
 Physiological measures (e.g., galvanic skin response)
A penile plethysmograph