Chapter 5 Flashcards
3 types of measures
self-report
physiological
observational
3 types of self reports
cognitive
affective/emotional
behavioral
self-reports can be formatted in what 2 ways?
rating scales or open ended questions
main difference between observational and physiological measures
obs: external and visible (behavior)
phys: internal and non-visible (bodily processes)
what guides us when choosing out type of measure?
theoretical interest & knowing that the 3 measured don’t always agree
researchers are encouraged to use how many types of measurement?
at least 2
properties of the real number system & definitions
identity: each # has a unique meaning
magnitude: #s have a correct order
equal intervals: dif. b/w units is equal anywhere on the scale
true zero: complete absence of the variable; NOT a quantitative value
example of a non-true zero
temperature 0 degrees Celsius
nominal variables
-no numerical properties
-doesnt make sense to do math
-can’t be ordered
-qualitatively distinct categories
orfinal variables
-can rank levels of a variable
-distances between consecutive values are not equal
-higher numbers represent more of the quality being measured
measurement
process of assigning values meaningfully to behavior, people, etc.
interval variables
-distances between consecutive values ARE EQUAL
-no true zero
-ex: temperature
ratio variables
-equal intervals
-true zero point/absolute zero point
-ex: time, length, weight, etc.
scales of measurement and their numerical properties
nominal - identity
ordinal - identity, magnitude
interval - identity, magnitude, equal intervals
ratio - identity, magnitude, equal intervals and true zero
considerations when choosing a scale of measurement
-info yielded
-stats
-ecological (do they correspond meaningfully)
reliability
how consistent the results of a measure are
observed score
true score plus error
true score
the score if the measure were perfect; not directly observable
sources of measurement error
-mood of participants
-stable attributes of participants (motivation, social anxiety, etc.)
-factors in the research setting
-characteristics of measure (ambiguity, test length, etc.)
-recording mistakes
if consistency is observed…
the measure is tapping into true scores
3 types of reliability
test-retest
interrater
internal
test-retest reliability
consistency in results every time a measure is used
what and when do we observe test-retest reliability for?
measure the SAME sample of people at at least 2 DIFFERENT points in time.
what measures does test-retest apply to?
self-report, obs. and phys.
what coefficient quantifies test-retest reliability?
correlation baby!
2 limitations of test-retest
-only meaningful for stable constructs
-responses can be artificially inflated if people remember their initial responses
what graph do we rely on for test-retest?
scatterplots
interrater reliability
consistency across observers for the same set of obs.
interrater reliability is most applicable to which measure?
obs.
interrater reliability for quantitative variables
correlation coefficient
interrater reliability for categorical variables
% agreement or kappa
is kappa usually smaller or larger than % agreement?
smaller
internal reliability/consistency
strength of correlation between items
what and when do we observe internal reliability for?
measure the same people with MULTIPLE ITEMS tapping into SAME CONSTRUCT
AIC
average inter-item correlation: average of all possible correlations between each other and the others
what AIC shows good internal reliability?
0.15-0.50
what cronbach’s coefficient alpha shows good internal reliability?
0.80-1
2 signs of low internal reliability
-more measurement error
-construct has more than one component
reliability is likely to increase if we increase the number of…
-observers
-observations
-occasions
5 types of measurement validity
face
content
criterion
convergent
discriminant/divergent
face validity
does the measure at the surface level (i.e. just looking at the term) appear to measure what it’s supposed to?
which type of validity may not be present but the measure can still actually be valid?
face validity
content validity
the measure contains all the domains that the theory says it should contain
criterion validity
is the measure related to a concrete behavioral outcome that it SHOULD be related to?
concurrent vs predictive criterion validity
concurrent: measure and outcome assessed at the SAME TIME
predictive: outcome assessed at a later time
what is criterion validity typically represented with?
r
convergent validity
does the measure strongly correlate to other measures of the same construct?
divergent validity
does the measure not correlate with other measures that it shouldn’t correlate with?
what type of relation is found in discriminant validity?
weak/zero relation
what type of relation is found in convergent validity?
strong positive or negative relation
a more concentrated scatterplot is convergent or divergent?
convergent
a measure cannot be _______ without _______
valid;reliability