Summary PSychometrics Literature

ESTIMATING & EVALUATING
CONVERGENT 1
& DISCRIMINANT VALIDITY EVIDENCE


Nomological Network: network of meaning around a
construct (need for belongingness, intimacy,
extraversion). Dictates a pattern of associations among
measures of the constructs




 Validation process: estimating degree to which test
scores actually show the predicted pattern of
associations.
 We look at the strength of the correlations not the
direction (-/+)

Methods to evaluate convergent and discriminant
validity
1. Focused Associations: focus on few highly relevant
criterion variables
 Verification of associations provides strong validity
evidence, but research failing to verify the associations
casts serious doubts on validity.
 e.g. sensitivity of parents, physical contact, quality of
interaction > highly relevant for attachment
2. Validity Generalization: evaluating test’s validity
coefficients across a large set of studies. (Do they
generalize to other settings, times, groups, situations etc.?)
 Is the relationship between physical contact and secure
attachment same for mothers and fathers, boys and girls?
 Reveal the level of validity across all of the smaller individual
validity studies (e.g. average validity correlation bt NEO-PI-R
Consc. and job performance)

, ESTIMATING & EVALUATING
CONVERGENT 2
& DISCRIMINANT VALIDITY EVIDENCE
 Reveal the variability in validity coefficients among the
smaller studies
 Reveal the sources of variability/ moderators that can
explain variability in validity coefficients (e.g. culture,
measurement method)