validity
VALIDITY DEFINITIONS AND TYPES
Validity: Defined as “the degree to which a procedure or assessment [does X, Y, or Z].”
Types of validity include:
Internal Validity: Concerns the extent to which a study establishes a trustworthy cause-and-effect relationship.
External Validity: Refers to how applicable the findings of a study are in the real world, also known as generalizability.
Construct Validity: Refers to the degree to which a test measures what it claims to measure.
Statistical Validity: Pertains to the appropriateness of the statistical analyses used and the conclusions drawn.
Two Subjective Ways to Assess Validity:
Face Validity: The degree to which a procedure or assessment appears effective in terms of its stated aims.
Content Validity: The degree to which a procedure or assessment represents all facets of a given construct.
Three Empirical Ways to Assess Validity:
Reliability: Refers to the consistency of scores or responses over time.
Types of Reliability:
Test-Retest Reliability: Achieved when people receive consistent scores upon retaking a test.
Internal Consistency Reliability: Indicates consistent responses on every item of a questionnaire.
Interrater Reliability: Consistency among raters when evaluating the same set of subjects.
Criterion Validity: The measure is correlated with a relevant behavioral outcome.
Predictive Validity: The degree to which a procedure or assessment predicts a future outcome.
Concurrent Validity: The degree to which a procedure or assessment aligns with another measure given at the same time.
Convergent Validity: The self-report measure is more strongly associated with self-report measures of similar constructs.
Discriminant Validity: The self-report measure is less strongly associated with self-report measures of dissimilar constructs.
FACE VALIDITY
Definition: Face validity refers to the degree to which a procedure or assessment appears to be effective in terms of its stated aims. It is a subjective and informal measure, serving as a good starting point but not comprehensive on its own.
Criticism: Low face validity can indicate inadequate measurement, while high face validity suggests that the assessment meets its aims.
Example: Measuring risk for heart disease using a Body Mass Index (BMI) assessment may differ in perceived validity, giving insights into the measure’s perceived truth in its evaluation.
CONTENT VALIDITY (LOGICAL VALIDITY)
Definition: The degree to which a measure covers all facets expected to be included based on the theoretical construct.
Essential Questions: Does your measure logically cover everything it should?
Examples for Workplace Satisfaction Scale:
On a scale of 1-10, how much do you like your job?
On a scale of 1-10, how likely are you to leave your job in the next 6 months?
Relative to other jobs you have held, is the management here better, worse, or about the same?
Possible facets to measure work satisfaction can include quality training, appropriate pay, collegiality, workplace safety, management style, commute, and benefits.
CRITERION VALIDITY
Definition: The degree to which a procedure or assessment performs against a separate set of criteria.
Types:
Predictive Criterion Validity: The degree to which an assessment predicts a future outcome.
Concurrent Criterion Validity: The degree to which an assessment aligns with another measure at the same time.
VALIDITY EXAMPLES
Predictive Validity Examples:
Whether SAT scores are indicative of future college success.
Study by Teramoto et al. (2018) examining physical attributes' correlation with defensive performance in NBA, showing significant positive correlations (r = 0.313-0.545).
NFL Combine's low predictive validity for gameday performance with the exception of sprint tests for running backs.
Concurrent Validity Examples:
Leadership Aptitude Test (LAT), where results correlate with supervisors' assessments of leadership abilities.
Self-Esteem tests measuring student ratings against teacher judgments, establishing the validity of new tests if found to correlate.
Nursing Competence Assessment (NCA), where the new tool’s results correlate with supervisor evaluations but possess low predictive validity.
CONVERGENT AND DISCRIMINANT VALIDITY
Convergent Validity: The extent to which different measures of the same construct correlate with each other, affirming their relationship.
Discriminant Validity: The extent to which measures of different constructs do not correlate, affirming their distinction.
Analogy: In a military context, cooperation among unit members should exist, whereas no cooperation should occur with enemy members, highlighting the essence of both types of validity in assessments.
Importance: To establish good construct validity, both convergent and discriminant validity are required.
CONSTRUCT VALIDITY MEASUREMENT
Measurement Method: Similar to reliability; involves correlations to assess validity.
Subjective Validity: Face and content validities should be evaluated informally, while predictive, concurrent, convergent, and discriminant validities require statistical evaluation.
RELIABILITY COEFFICIENTS
Interpretation of Correlation Coefficients:
A value of $0.00$ indicates no relationship.
Values $0.01$ to $0.24$ suggest weak relationships.
Values $0.25$ to $0.49$ suggest moderate relationships.
Values $0.50$ to $0.74$ indicate strong relationships.
Values $0.75$ to $0.99$ show very strong relationships, while $1.00$ is perfect reliability.
Common Standards:
Interrater reliability: $0.85$
Discriminant reliability: > $0.30$
Strong correlations: > $0.7$
Weak correlations: < $0.50$
SCALES OF MEASUREMENT
Qualitative Variables:
Nominal: Categorical, e.g., color, gender.
Quantitative Variables:
Ordinal: Rank order matters; example: placing in a competition.
Interval: Both rank order and intervals matter; examples include temperature in Celsius/Fahrenheit.
Ratio: Includes all features with a true zero; examples include the number of vehicles owned.
SUMMARY OF VALIDITY TYPES AND IMPLICATIONS
Internal Validity: Ensures strong cause-and-effect relationship without confounding.
External Validity: Ensures results generalizable in real-world settings.
Assessing Validity: Utilizing reliability measures supplemented with correlation assessments ensures integrity in research measurements, thereby enhancing the validity of outcomes and applications.