Validity and Reliability

Get Started. It's Free
or sign up with your email address
Rocket clouds
Validity and Reliability by Mind Map: Validity and Reliability

1. Content Validity Evidence

1.1. Inspect the test questions to see whether they correspond to what the user decides should be covered by the test. The problem with this evidence is that the test can look valid but it might not be measuring what it is intended to measure.

2. Criterion-Related Validity Evidence

2.1. Scores from a test are correlated with an external criterion.

2.1.1. Concurrent Criterion-Related Validity Evidence

2.1.1.1. Measures that can be administered at the same time as the measure to be validated. Determined by administering the new test and the established test. This also helps give validity of the new test.

2.1.2. Predictive Validity Evidence

2.1.2.1. How well the test predicts some of the examines future behaviors. This is determined by testing the subjects, measuring the subjects on the content of the test to predict the future outcome after a period of time has passed. Example Test: aptitude, personality, benchmark

3. Construct Validity Evidence

3.1. Relationship to other information corresponds with some theory. Logical explanation or rationale for the interrelationships among a set of variables. Information that lets you know whether the results correspond with what you expected from test.

4. Test-Retest or Stability

4.1. The problem with this method is there is usually some memory involved between the two test.

4.2. The test is given twice within a given time and the correlation between the two scores is decided.

5. Alternate Forms or Equivalence

5.1. The advantage is that this eliminates the memorization possiblility.

5.2. Both forms are given and the scores between the two test are decided.

5.3. The problem with this method is knowing that you have TWO good tests.

6. Internal Consistency

6.1. People should get similar items correct. Items ought to be correlated with each other and the test should be internally consistent,.

6.1.1. Split-half Methods

6.1.1.1. The advantage is that only one test has to be given

6.1.1.2. Split the test into half and then decided the correlation between the two halves.

6.1.2. Kuder-Richardson Methods

6.1.2.1. Measure the extent to which items on one form have as much in common as the items on the equivalent form. Fairly consistent method.