1. New node
2. Methods of Validity
2.1. Content Validity
2.1.1. Easiest way to show validity
2.1.2. Helps specify what should be included in a test
2.1.3. Gives information on if a test looks valid, but not on if test is age appropriate or is properly constructed. Usually based on judgement.
2.1.4. Reveals if test matches classroom or learning objectives.
2.2. Criterion-Related Validity-test scores are correlated to an external criterion
2.2.1. Concurrent Criterion-related validity-gives a numerical value to show validity. Used to determine if more than one test measuring the same thing can both yield the same results. If both give the same result then one may be used over the other based on affordability and time constraints.
2.2.2. Predictive Validity- used to determine how a future behavior can be predicted through a test.This is determined by administering the test and then a short time later measuring the behavior the test is supposed to predict. Examples of tests include SAT, which is used to determine who is admitted into college.
2.3. Construct Validity- validity of test is based on theories. Test results should correspond with what you would expect them to.
3. Methods of Reliability
3.1. Test-Retest or Stability
3.1.1. Test is given twice and correlation is determined between them
3.1.2. Reliability cannot fully be determined due to factors such as memory or experience affecting re-test.
3.2. Alternate Forms or Equivalence
3.2.1. Equivalent forms of tests can be used to determine reliability of results.
3.2.2. Used to provide test security. Different versions of the same test may be given as opposed to coming up with two reliable tests.
3.3. Internal Consistency
3.3.1. Split-half method-single test is split in half and scores on both halves are used to determine correlation. Does not require two tests or more than one test administration. Although shorter it underestimates the reliability of the full test.
3.3.2. Kuder-Richardson methods - reliability is measured through how much items on one form of a test have in common with items on an equivalent form of the test. Reliability measures higher than the test-retest and alternatate forms.