I. Introduction
 
 

Define validity

Define construct and relate to validity

II. Types of Validity
 

A. Construct validity. This is the broadest and really encompasses all the others. We can never really prove that -
- we just get evidence supporting it. What we are looking for is evidence that supports our theoretical expectations.
 
 
 
 

Different types of evidence that all converge to look at construct validity.

(1) Evidence that the test is homogenous and has internal consistency.
(2) Evidence that the test measurement changes with age if the construct itself changes with age.
(3) Evidence of pre-test/post-test changes.
(4) Evidence from distinct groups.
(5) Convergent evidence
(6) Discriminant evidence the multi-trait multi-method matrix
(7) Factor analysis.

The other kinds of validity are those that contribute to obtaining one or the other of these types of evidence and thus to
supporting construct validity. They are:

B. Face Validity - does the test look like it measures what it says it measures
 

C. Content Validity - does the measure you have adequately sample the whole range of content that is included in this
construct?
 

D. Criterion Validity - How adequately a test score can be used to infer an individual's most probable standing on an
accepted criterion.
 

Since the criterion itself plays a part in limiting the validity coefficient let us look at it.
 

What are the characteristics of a good criterion?
 

There are two forms of criterion validity.
 
 

a. concurrent validity
b. predictive validity.

USES OF PREDICTIVE VALIDITY
 

1. To make individual predictions:
 

2. To help make decisions regarding selection & placement.
 
 

a. incremental validity
b. expectancy data
          expectancy table


selection efficiency analysis (Decision theory).
    Hit rate
    Miss rate
    Base rate
    Selection rate
            Taylor Russell tables

VALIDITY AND TEST BIAS

BIAS

Three ways in which you can see bias:

1. If the test itself is biased.
 

2. If there is a bias in rating.
 

a) leniency error
b) severity errors
c) central tendency error

3. Halo effect bias.
 

FAIRNESS
 

Fairness is a judgment that's made every time a test is used with a particular group or individual.

BE SURE TO STUDY THE CLOSE-UP ON FACTOR ANALYSIS IN THE CHAPTER.
 

BE ABLE TO DISCUSS THE ISSUES INVOLVED IN ADJUSTING SCORES FOR DIFFERENT GROUPS