3: Construct Valdiity

Question 1

Construct validity

Answer 1

• depends on conceptual variable
• reliability necessary but not sufficient

Evaluating abstract constructs;
Subjective
•face
•content

Empirical
•criterion (concurrent or predictive)
•convergent and discriminant

Question 2

Face validity (subjective)

Answer 2

Looks like a plausible measure of conceptual variable

Question 3

Content validity (subjective)

Answer 3

It includes all important parts of construct (all parts theory says it should contain)
-does measure sample of knowledge, skill and or behaviors.

Question 4

Criterion

Answer 4

Weather measure is related to a relevant concrete outcome that should be related to

• do subjects that are high or low behave as predicted from theory
• do values correlate with values of some criterion outcome

Question 5

Correlation coefficient (criterion)

Answer 5

-evidence for criterion when outcome is quantitative (interval or ratio)

Question 6

Known groups (criterion)

Answer 6

Evidence for criterion when criterion outcome is categorical

Question 7

Concurrent criterion validity

Answer 7

-criterion data collected at about the same time

Question 8

Predictive criterion validity

Answer 8

Outcomes occur some time in the future

-measure thus predicts future behavior or outcomes that should be associated with construct being measured

Question 9

Convergent

Answer 9

Measured correlated strongly with other validated measures of the same
Construct

Question 10

Discriminant

Answer 10

Measure correlates less strongly with validated measures of differnt constructs

Question 11

Established measures v new measures

Answer 11

Established: have already been validated

• don’t need to collect data assessing reliability or validity
• refer to previous research for reliability and validity

New: must be validated

• takes substantial time and effort
• evaluate accuracy, sensitivity, range effects if applicable