PSYC 111 Lecture Notes - Lecture 19: Predictive Validity, Content Validity, Standardized Test
Document Summary
Test construction: because scores become meaningful only when they can be compared with others" performances, they must be defined relative to a pretested group, a process called standardization. Group must be representative of those who will be taking the test in the future. Standardized test results typically form a normal distribution, a bell-shaped pattern of scores that forms the normal curve. Most scores cluster around the average and increasingly fewer are distributed at the extremes: flynn effect. Past several decades the average sore has raised, a phenomenon known as the flynn effect. New zealand researcher james flynn first calculated its magnitude. The average person"s intelligence test score is 1920 was, by today"s standard, only a 76. This supports the need for ongoing standardization/updating normative samples: reliability. The extent to which a test yields consistent scores. Consistency may be assessed by comparing scores on two halves of the test (split- half), on alternative forms, or on retesting (test-retest)