Standardized tests, which are academic examinations administered to large groups of people, are described as "standardized" largely because each individual's test is graded in the same way. In many cases, such tests are graded by machine, with test takers filling out answers on "bubble sheets" that are run through a computer for grading. In other cases, such as standardized writing exams, a human grader is required, as computers aren't yet capable of analyzing writing skills. In this situation, many graders often are assembled to handle the work, and one test might be read by several people to help ensure objectivity. It is nearly impossible to remove subjectivity or arrive at a grading process consensus when humans are involved, however, and as such, the practice of having humans grade standardized exams is somewhat controversial.
This standardization in grading standardized tests is highly important, as such tests are often highly competitive. Standardized tests are, for instance, examined and compared in the determination of college, law school, medical school, and graduate school admissions. Conformity in grading standardized tests is absolutely essential in order to ensure that such determinations are made fairly and that one individual is not chosen over another simply because of bias or error in the grading process.
Machines can be used for grading standardized tests composed of multiple choice or true-false questions. Many standardized tests are, therefore, composed of such question types because machine grading has a relatively low rate of error and is free from bias. Additionally, the process of grading standardized tests can be completed with great rapidity when all of the tests can be fed through a machine. This is very important for tests that are administered throughout an entire country, as the workload can be far too great for human graders to handle effectively. Such tests are generally administered on "bubble sheets," on which test takers fill in a labeled "bubble" on the answer page in order to select their answers.
Not all skills can be tested through the use of bubble sheets, however. Writing skills and the ability to analyze and respond to issues with competence and depth often requires writing or detailed calculation that cannot be examined by a computer. Grading standardized tests with written elements requires the presence of a human grader to read and make judgments on the test taker's work. In some cases, several graders read over each piece of written work and compare scores in order to grant some level of objectivity to the grading process. In other cases, this is not possible.
The use of human intervention in grading standardized tests is somewhat controversial. People with different levels of expertise and different opinions are often unable to come to a consensus on how written work is to be graded. Attempts at remedying these problems include hiring graders only from within relevant fields and requiring graders to undergo extensive training. Biases, however, cannot be completely removed in a single training session, and different graders have been shown to grade the same pieces of writing in vastly different ways. Other factors such as how much the graders are paid and how much experience a given grader has may also bias the process of grading standardized tests.