Study Guide 11: Construct Related Evidence Flashcards Preview

EPSE 528 > Study Guide 11: Construct Related Evidence > Flashcards

Flashcards in Study Guide 11: Construct Related Evidence Deck (12):

Campbell & Fiske (1959):

developed multitrait-multimethod matrix


Convergent evidence

degree to which test scores are correlated with tests of related constructs


Cronbach & Meehl (1955)

developed the idea of nomological network to give meaning and structure to constructs


Discriminant evidence

degree to which test scores are uncorrelated with tests of unrelated constructs


Known groups validity or contrasted group studies

validity is determined by the degree to which an instrument/test can demonstrate different scores for groups known to vary (or not vary) on the variables being measured.


Method effects or common method variance:

Occurs when HTMM correlation > HTHM correlation and shows that use of the same method will inflate the correlation effect.


Multitrait-multimethod matrix (MTMM):

Provides guidelines for evaluating construct validity.
By considering the effects of trait and method variance on correlations among measures, researchers can gauge quality of convergent and discriminant validity evidence.


Nomological network

interconnection between a construct and other related constructs (correlations), embedded in theoretical context


Quantifying construct validity (QCV):

quantifying the degree of “fit” between theoretical predictions for convergent and discriminant correlations and actual sets of correlations obtained.


R2 (or r2):

squared correlation – sometimes interpreted as being the proportion of variance in one variable explained by another variable. Criticized for sometimes being statistically incorrect, non-intuitive metric, and minimizing the importance and magnitude of correlations.


Validity coefficient

correlation between test scores and highly relevant variables


Validity generalization

process of evaluating a test’s validity coefficients across a large set of studies. It is intended to evaluate predictive utility of test scores across settings, times, situations, etc. Provides 3 pieces of info: general level of predictive validity, degree of variability among individual studies, and sources of variability.