CONSTRUCT-RELATED EVIDENCE of validity is most relevant in interpretation of standardized tests. Construct-related evidence suggests the extent to which the results can be interpreted as psychological constructs.
(That definition may not be especially helpful!)
Let me try again. Construct validity provides information about what a test is "really measuring", regardless of the name on its cover.
The primary tool in gathering construct-related evidence of validity is a mathematical procedure called factor analysis. The correlations between various tests are analyzed to determine how many different things (factors) appear to actually be measured with the test.
_____________________________________________________________________
back to measurement primer menu
next screen_____________________________________________________________________
For example, a test battery is administered which includes four separate tests: reading comprehension, vocabulary, grammatical usage, and spelling. There is a score for each test. It would appear that four things are being measured.
The table below has the correlations between each combination of these tests.
vocab gramm use spell
-----------------------------
rdg comprehension .94 .91 .10
vocabulary .92 .04
grammatical usage .15
----------------------------------------------------
This table shows high correlations (.90+) among the combinations of the reading comprehension, vocabulary, and grammatical usage tests. Each of the three has a lower correlation with spelling.
_____________________________________________________________________
back to measurement primer menu
next screen_____________________________________________________________________
From these data there would appear to be only two, not four, different things being measured with this test battery.
One construct could be called verbal facility and includes the reading, vocabulary, and grammar tests.
Spelling appears to be measuring a different construct.
vocab gramm use spell
-----------------------------
rdg comprehension .94 .91 .10
vocabulary .92 .04
grammatical usage .15
----------------------------------------------------
_____________________________________________________________________
back to measurement primer menu
next screen_____________________________________________________________________
With standardized tests it will often be a good idea for you to look at the construct validity evidence in the test manual before you make inferences about a person's characteristics based on the test scores.
For example, a short academic aptitude test administered along with a standardized achievement test battery may not provide appropriate evidence for you to identify students whose achievement is inconsistent with their ability.
You may remember from the section on test norms that the fact an aptitude test score is a deviation IQ and the achievement test score is a percentile does not mean they are measuring different things. Any of the converted scores could be used with any of the tests.
Often, the construct evidence for short aptitude tests indicates that they are actually measuring the same thing as the achievement test. Using contrasts between scores on one and scores on the other to classify someone as an "underachiever" is not a wise inference.
_____________________________________________________________________
back to measurement primer menu
next screen_____________________________________________________________________
Which type of validity would be more important in a test designed to measure creativity?
a. construct-related
b. content-related
c. criterion-related
_____________________________________________________________________
_____________________________________________________________________
Which type of validity would be more important in a test designed to select students for admission to college?
a. construct-related
b. content-related
c.. criterion-related
_____________________________________________________________________
_____________________________________________________________________
YOU ARE CORRECT.
In this instance, you would look for evidence that this test is measuring an underlying construct of creativity.
It is also sometimes important to have more than one type of validity evidence. For example, in this case, you might also want to look for criterion-related evidence that the test score actually predicts creative behavior.
_____________________________________________________________________
back to measurement primer menu
next question_____________________________________________________________________
No, in this instance, you would probably want to first look for evidence that this test is measuring an underlying construct of creativity.
It is also sometimes important to have more than one type of validity evidence. For example, in this case, you might also want to look for criterion-related evidence that the test score actually predicts creative behavior.
_____________________________________________________________________
back to measurement primer menu
next question_____________________________________________________________________
YOU ARE CORRECT!.
The purpose of an admission test is to predict later performance. Criterion-related evidence would be most important in this case.
_____________________________________________________________________
back to measurement primer menu
_____________________________________________________________________
No, the purpose of an admission test is to predict later performance.
Criterion-related evidence would be most important in this case.
_____________________________________________________________________
back to measurement primer menu