MSR Intro
MSR Intro
(Psychological measurement)
Introduction to Psychological
Measurement
Scales
Well use the terms scale to mean a multi-item instrument
designed to represent the amount or kind of a specific attribute
for a specific individual
Population:
Domain:
The type of information we want to measure with our scale.
We sample items to represent the domain.
Construct value:
The individuals amount or kind of the attribute or
characteristic we are trying to measure.
Variable:
The individuals measured score or code
Reliability (Consistency)
You also have to pay attention to this when selecting scales for
research is the scale designed to give you the kind of
measure you want (equidiscriminating vs. classificatory) for
your target population????
Standardization
Administration -- given the same way every time
who administers the instrument
specific instructions, order of items, timing, etc.
Varies greatly -- multiple-choice classroom test (hand it out)
-- Intelligence test -- 100+ pages of how to in
manual -- about 1/2 semester in Psych 955
Reliability (Consistency)
Inter-rater or Inter-scorer reliability
can multiple raters produce the same score for a given test
(assumes standardization)
Internal reliability -- agreement among test items
split-half reliability -- randomly split into two tests & correlate
Chronbachs
Psychometric Sampling
from a statistical or research design perspective sampling
usually refers to the selection of some set of people from which
data will be collected, for the purposes of representing what the
results would be if data were collected from the entire
population of people in which the researcher is interested
from a psychometric perspective sampling is a broader issue,
with three dimensions
sampling participants (respondents) to represent the desired
population of individuals
sampling attributes to represent some desired domain of
characteristics or behaviors
sampling stimuli (may be other people or even themselves)
to represent the desired domain of objects
respondents give us values that represent attributes of stimuli
Examples
ul
i
St
im
People
3-way sampling
Attributes
Interrelationship - Evaluating one type reliability requires assuming the other type of
reliability
another
kinds of items
Survey Items
Scale Items
Test Items
b. Speed up
disagree 1 2 3 4 5 agree
1 2 3 4 5
1 2 3 4 5
value on item
Important terms
linear
monotonic vs. nonmonotonic
deterministic vs. probabilistic
specified vs. unspecified
distribution form
value on attribute
etc...
etc...
Deterministic Model
each item perfectly discriminates those who are above vs. below
that extent of the target attribute
so, a properly chosen item set will allow specific identification of
where each respondent is along the attribute continuum
a respondent answers no to all items less difficult than a
specific value, and yes to all items more difficult than
that value
1.0
0
Attribute value
1.0
4 items of
different
difficulties
0
Attribute value
Nonmonotone Model
each item is maximally responded to by respondents at a
certain position on the attribute continuum
both those above and below that position respond with lower
probability
Item 2
Examples of Item 2
Item 1
Item 3
Politically speaking, I am
a moderate
Higher
Item Response
great item
Lower
common items
bad items
Lower values
Higher Values
Attribute of Interest
sum of good
items