CRITERION-RELATED EVIDENCE of validity is typically in the form of a correlation coefficient (see statistical primer).
The evidence can also be in the form of an experience/expectancy table.
Criterion-related validity evidence indicates the relationship between scores on some test (or other observational procedure) and an outside CRITERION.
_____________________________________________________________________
back to measurement primer menu
next screen_____________________________________________________________________
When the relationship is "future", this evidence is called PREDICTIVE VALIDITY.
An example would be the relationship between an ACT test administered during the senior year in high school and grades during the first year of college.
When data on both test and criterion are gathered at essentially the same time, this evidence is called CONCURRENT VALIDITY.
An example would be determining the relationship between a short screening test and a longer, established test of the same characteristic.
_____________________________________________________________________
back to measurement primer menu
next screenThere are as many different sources of criterion validity evidence as there are criterion which could be correlated with the test score.
For example, your score on the first test in one of your college classes could be correlated with your score on the second test, your score on the final exam, your grade in the class, your height, your g.p.a., and so forth.
You use past examples of criterion-related evidence of validity as a tool to determine what kinds of inferences (interpretations) would be appropriate.
You might find, for instance, that correlation between score on the first test you administered and score on the second test you administered was .75 while correlation between score on the first test and the final grade was only .18.
The next year (assuming that tests and students are comparable) from results on the first test you could make a reasonable prediction (inference) about the student's performance on the second test. You wouldn't want to use the first test score, however, to try to predict the final grade.
back to measurement primer menu
next screenCriterion-related evidence is often in the form of a correlation coefficient. The table below is another way to present the information.
midterm grade in course
test F D C B A
-------------------------------
35 to 40 0 0 0 2 5
30 to 34 0 0 1 4 2
25 to 29 0 1 3 3 1
20 to 24 1 1 1 0 0
This table shows a strong relationship between midterm test and final grade. Of those who scored between 30 and 34 on the midterm, 2 made an A in the course, 4 made a B, and so forth. When a table like this is used to describe this class, it is called an EXPERIENCE TABLE.
If it is used at midterm in some future class to predict final grades, it is called an EXPECTANCY TABLE.
back to measurement primer menu
next screen_____________________________________________________________________
A brief intelligence test screening instrument (15-2O minutes) is administered along with the Wechsler Intelligence Scale for Children-IV (one and one-half hours). The correlation coefficient is calculated between the deviation IQ scores produced by the two tests. This is:
a. concurrent criterion validity
b. predictive criterion validity
c. concurrent content validity
d. predictive content validity
_____________________________________________________________________
_____________________________________________________________________
YOU ARE CORRECT.
The evidence is CRITERION because this is the relationship between a test and an outside criterion (WISC-IV). It is CONCURRENT because the test are administered at essentially the same time.
If there had been a significant time gap between administration of the screening test and administration of the WISC-IV, the evidence would be have been predictive.
(I don't think there is any such thing as concurrent content validity or predictive content validity).
_____________________________________________________________________
back to measurement primer menu
next screen_____________________________________________________________________
No, the correct answer is concurrent criterion validity.
The evidence is CRITERION because this is the relationship between a test and an outside criterion (WISC-IV). It is CONCURRENT because the test are administered at essentially the same time.
If there had been a significant time gap between administration of the screening test and administration of the WISC-IV, the evidence would be have been predictive.
(I don't think there is any such thing as concurrent content validity or predictive content validity).
_____________________________________________________________________
back to measurement primer menu
next screen