According to the University of California--Davis, a standardized test is reliable when its results are stable over time. That is, when the same person takes the same test on two separate occasions, the results should be somewhat similar in order for a test to be reliable.
Schools often use standardized tests to determine funding, curriculum choices, and as a measure of students' understanding of material. Other organizations use standardized tests to determine a person's intelligence or personality type. Knowing that a test is reliable can help teachers and administrators make appropriate choices based on a test's result.
The psychology department of Valdosta State University reports that standardized tests are measured for their reliability, as well as their validity. According to VSU, "Validity, very generally, is concerned with how well a test measures what it is supposed to (e.g., accuracy), while reliability is the extent to which a test consistently measures what it is measuring."
Because reliability is an important aspect of standardized tests, test makers can work to increase reliability when they're constructing tests. Special Connections, an education publication of the University of Kansas, recommends increasing reliability simply by having more questions on the test. Having more questions decreases the impact of one or two careless mistakes that don't reflect the taker's actual knowledge of the content.
In addition, most standardized tests are objective, and this increases reliability, as subjective answers are more open to the scorer's interpretation.
To determine if a standardized test is reliable, the University of California--Davis suggests testing the reliability of the standardized test. This can be done in three ways.
The test-retest method is where the same group of people is given the same test on two separate occasions, and the test is considered reliable if the results are similar. In a split-half testing, the answers to the test are split in half, and if each half has approximately the same number correct and incorrect, then the test is considered reliable. Finally, with the alternate forms method, two versions of the test are made, and they're given to the same people on separate occasions. If the results are similar, then the test is considered reliable.