Page 359 - Understanding Psychology

Figure 13.2 Judging Reliability
High
High
On the left, the test scores obtained by seven individuals are ordered on a scale. On the right, the cor- responding scores on a second version of
Low
High
Low
High Reliability
Low the same test, given at a later time, are
High
Low
ordered. In the upper diagram, the two sets of scores correspond very closely—meaning the test is highly reli- able. What is meant by saying a test is reliable?
Low Reliability
In checking tests for reliability, psycholo- gists try to prevent variables from influencing a person’s score. All kinds of irrelevant matters can interfere with a test. No test can screen out all interferences, but a highly reliable test can eliminate a good part of them.
TEST VALIDITY
A test may be reliable but still not valid. Validity is the ability of a test to measure what it is intended to measure (see Figure 13.3). For example, a test that consists pri- marily of Spanish vocabulary lists will not measure ability for engineering. A history test will not measure general learning ability. A test you take in physical education class may not measure your knowledge of gram- mar, or a math test that asks questions that were not covered in class does not measure what you learned in class.
Determining the validity of a test is more complex than assessing its reliability. One of the chief methods for measuring validity is to find out how well a test predicts performance—its predictive validity. For example, a group of psychologists designs a test to measure management ability. They ask questions about management systems, attitudes toward employees, and other relevant information. Will the people who score high on this test really make good managers?
Suppose the test makers decide that a good way to check the validity of the test is to find out how much a manager’s staff improves in pro- ductivity in one year. If the staffs of those equally skilled managers who scored high on the test improve more than the staffs of those managers who scored low on the test, the test may be considered valid. Corporations may then adopt it as one tool to use in deciding whom to hire as managers, assuming the test is also valid for their situations.
What if managers who are good at raising productivity are poor at decision making? It may be that this test measures talent for improving productivity, not general management ability. This is the kind of diffi- culty psychologists encounter in trying to assess the validity of a test. As the example shows, nothing can be said about a test’s validity unless the purpose of the test is absolutely clear.
STANDARDIZATION
Tests must be standardized. Standardization refers to two things. First, standardized tests must be administered and scored the same way every time. Test administrators are trained to follow the same procedures and
validity: the ability of a test to measure what it is intended to measure
Chapter
13 /
Psychological Testing 345
Test Score Test Score
Retest Score Retest Score