Instrumentos de evaluación

Abstract

Two factors are important in the development of good tests: reliability and validity. Generally speaking, more extensive tests are more reliable than others, i.e., reliable tests will go on using scores as long as non reliable tests exist. This way, the test is reliable. All tests are less-than-perfect indicators of the qualities or abilities being measured. Errors may occur in all test settings. Thus, a reliable test may also be defined as a test with a small measurement standard error. Furthermore, test items are not easy to elaborate. Interdependent questions decrease the accuracy of a test, and hence its validity. Most smart testers are aware that categorical answers are usually wrong. If only one answer is right and two others are equal, then these two should be wrong; such items are relatively easy to elaborate, since they are mere declarative statements. Concurrent and predictive validity: If a new test is validated by comparison with an existing parameter, it has con current validity. Tests designed to predict their own results have predictive validity. Total validity = content validity + criterion validity + construct validity. Thus, the validity of a instrument for measurement is assessed based on three types of evidence. Validity and reliability both measure a degree, not an absolute value, i.e., a test may have a higher or lower degree of reliability or validity.