Why be concerned with psychometrics?
1. To identify boundaries of appropriate use
-To avert misuse of abuse of measures
2. To know constraints on interpretation of results
-Limits of meaning
3. To understand the construct(s) tapped by the measure.
4. TO ensure that patients can benefit from assessment practice.
-A measure of association.
-Close to 0.0= bad, close to 1.0= good.
-The replicability of a score
-The consistency or stability of a score across measurements
-The proportion of true-score variance in a fallible score
-Sources of Error and the Sampling universes: -
1. Item content…
-behaviour sampled by items -alternative form
-domains of behaviour -split-half
2. Assessment episode
3. Raters and Interviewers
-constant errors -inter-rater
-the meaningfulness of a score
-the degree to which a score corresponds to the construct that the
measure is assumed to quantify
Methodologies for estimating validity:
1. Concurrent validity
-correlation between a score and a criterion score
2. Predictive validity
-correlation between a score and a criterion outcome
3. Content validity -the items represent the target universe
4. Face validity
-Construct validity involved a simple question
-What is the construct that accounts for score variance?
-But it is not easily demonstrated!
-merges with theoretical and empirical research
-involves formulating and testing hypotheses
1. A construct is defined implicitly, by a network of propositions.
*Intrinsic Motivation Inventory for Schizophrenia Research
-I enjoyed doing this activity very much
-This activity was fun to do
-I think this is a boring activity
-I put a lot of effort into this
-I tried very hard on this activity
-It was important to me to do well at this task