For the diagnosis to be termed accurate, the features identified have to be same every time if the test is to be termed reliable. This is measured by a correlation between the two test outcomes over all the participants. This method expresses the generalisability of the test scores over different test occasions by assessing stability of the test outcomes (Hempel, 2005).
Hempel (2005) explains that this method of testing reliability varies with the type of tests, i.e. ability or traits and attitudes or states assessment. In both cases, repetitive responses based on previous memory is a high possibility thereby giving incorrect measures of reliability. Secondly, the traits tend to change with time, and this would depict low reliability measure, which is again incorrect (Garb, 2002). In such cases, the time period between the tests needs to be recorded appropriately in order to obtain accurate reliability measure. In case of validity, the test should be able to measure what is intended to be measured with minimum or no possibilities of change in behaviors. Reliability of the test decides the amount of correction and incorrect decisions that are made as an outcome of the test. Therefore, reliability is necessary for confirming validity. At the same time, validity of results obtained from this method is higher when the time gap is minimal; however, this minimum time varies with situation or behavior being