Validity refers to the extent that a test measures what should be measured. It said the problem is whether a measurement (e.g performance measurement) to measure what should be measured. If it does not measure what is supposed to measure the problem will arise. A test is said to be valid if the test is reliable and what measures should be measured. Validity can be divided into: the validity of the face, the validity of content, and Criterion Validity The validity of the idea.
a. Face Validity
Type of validity that focuses on whether the items in the questionnaire is clear and understood by the subjects. This refer to it measure what is measured with the same name.
b. Content Validity
Refers to the extent to which the test covers topics in a field of study. They measure the concept adequately studied. For example in the evaluation of employee performance; whether the questions raised in line to get information about the performance of the entire staff.
c. Criterion Validity
A test is said to have validity if the test criteria used correlated with other tests using different measurement tools. In other words, we use as test estimator for a behaviour or attitude. For example, if a person has undergone a driving test and was confirmed pass, then we can estimate that he will be able to drive well and carefully.
a. Concurrent Validity
Used to measure a phenomenon at present. For example, in a factory, a measure that they have reviewed the physical training every morning is get out their higher productivity than those who do not. This means that there is a relationship between exercise performed in the productivity of production. Another example is the validity of the line, whether a student who obtained good results in the examination of a good tingkahlakunya. If both are parallel, then our test is valid.
b. Predictive Validity
Predictive validity (predictive) refers to the ability of the measurement tool to predict an event in the future. For example, the relationship between STPM CGPA of students at the university. If someone is successful in the foreseeable STPM he will succeed at university. Some researchers also say that the matriculation results is more accurate to predict success at university.
d. Construct Validity
The construct validity is essential in testing the theory. It is associated with the test have to do with the other test results obtained from a theory. It is related to the theoretical or empirical scientific observation. Researchers not only confirmed the test but also try to verify the theory in the test. The validity of the idea can be divided into two, namely:
i. Convergent validity
Validity refers to the verification of a measuring instrument can measure whether the concept as described in the theory. Does measurement measure the concept as described in theory.
ii. Discriminant Validity
This validity refers to the verification of a device with a means of measuring the correlation between the instruments. Are survey instruments have a low correlation with a variable that has no relevance to it?
Reliability refers to the extent of tests or procedures such as repeated tests will give the same results. In general, reliability refers to consistency in the measurement of a set of data obtained. It is the level at which no error in the measurement and the same test with a high degree of consistency. For example in the measurement of student achievement, we expect the same performance if we take the same test at different times. In short, we want to create the impression that based on tests we provide, we can predict students' performance in the same test but in the future or in a different environment.