Lecture 9 Flashcards
Construct validity
To what extent do the test scores measure the latent construct they should measure?
Criterion validity
how well a test predicts or correlates with an external standard outcome (the criterion)
How is criterion validity different than convergent validity?
1-Criterion focuses prediction rather than the latent variable.
2-Convergent focuses on latent variable.
3-Convergent (correlation analysis )
4-Criterion ( regression analysis)
Types of cirterion validity
1-Concurrent: latent and criterium measured simultaneously
2-Predictive: criterium measured after latent
3-Postdictive: criterium is after latent variable ( test measuring preoccured trauma)
What are methods to evaluate Convergent and discriminant validity?
1-Focused association
2-Nomological network
3-Multi-trait method design
Nomological network
1- Relationship between a construct and a wide range of other constructs.
2- extended nomological network
3-Method used: largely visual inspection subjective evaluation of convergent and discriminant validity.
4-Not a formal method.
5- Correlation with the expected theory?
Multitrait-Multimethod
Multiple traits measured using multiple methods.
Why do we use multitraitmethod?
1-Looks at convergent and divergent validity in one design
2-Gets insights into method effects
Method and traits
Method: Refers to raters or questionnaires
Trait: latent variable
Method effect:
inflate correlations between different traits measured by the same method, reducing discriminant validity. ( Mom rating 2 different traits will inflate their correlation)
Response bias
1-Unintended psychological process related to taking the test that interferes with intended item response.
Definition: tendencies of participants to respond inaccurately or falsely to questions during self-reports or interviews.
