None of activity in this world which can be separated from
the activity measurements. The success of a program can be known through a measurement. The development of science and technology cant be separated from the activity measurements. The studies were carried out in all areas always involves measuring, both quantitatively and qualitatively. Therefore, measurement plays an important role, both for the development of science and technology as well as for the presentation of information for policy makers. Basically measurement is the determination of the number of activities for an object systematically. Determination of this figure is an attempt to describe the characteristics of an object. The ability of a person in a particular field is expressed with numbers. In determining individual characteristics, measurements taken should wherever possible contain small errors. Measurement errors that occur in the natural sciences is simpler than the measurement error in the social sciences. Errors in the natural sciences are mostly caused by the measuring instrument, while the measurement error in the social sciences could be caused by the measuring instrument, how to measure, and the state of the object being measured (Djemari Mardapi, 2008). Quantitative measurement that can be divided into three kinds, namely: (1) The measurements taken not to test something, such as measurements made by a tailor on the length of the arms, legs, shoulder width, waist size and others. (2) Measurements were carried out to test something, like measurement to test the durability of motorcycle engines, measurements to test the durability of incandescent lamps, and others. (3) measurements to assess done by testing something, such as measuring the learning progress of learners in order to fill the grades were done by testing them in the form of achievement
test. Measurement of the third type is what is known in the world
of education (Anas Sudiyono, 1996). Things that includes evaluation of learning outcomes include measuring instruments used, how to use, assessment, and evaluation. Measuring instruments used can be a chore, quizzes, midterms (UTS), and final exams (UAS). In principle, the measuring instruments used must have proof of validity (validity) and reliability (reliability) is high. Validity or the validity of the measuring instrument can be seen from the constructs of measuring instruments, which measure something that is planned to be measured. According to the theory of measurement, the substance being measured must be one-dimensional. Aspects of language, writing neatness is not suspended or taken into account when measuring the goal is to determine the ability of learners in certain subjects. Construction measuring devices can be examined on the material aspect, itemwriting techniques, and the language used. Experts in the field or colleagues is a good reviewers to provide feedback about the quality of measuring instruments used include test. The validity of the measuring instrument can also be seen from the grating measuring instrument. This lattice contains material tested, the shape and number of questions, the level of thinking involved, weight problems, and ways of scoring. Good lattice is representing teaching materials. For the subjects tested were selected based on the following criteria: (1) The essential subject matter, (2) has a value of application, (3) continuous, (4) needed to study other subjects. Another thing that is important is the length of time available to do the exam. Some argue, grilles should be delivered to the learner. The measurement results should have errors as small as possible. The error rate is related to the reliability of measuring instruments. Good measuring tool gives constant results when
used repeatedly, as long as the ability is measured not changed.
There are measurement errors are random and there are systematic. Random error due to the current situation of the exam, physical-mental condition that is measured and the measure varies. Mental conditions including a person's emotions can be varied, and assumed to be random variations. This is to make it easier to estimate a person's ability. Systematic errors caused by the measuring instrument, which is measured, and that measure. There are teachers who tend to create test questions that are too easy or difficult, so that the measurement results can underestimate or overestimate of the actual ability. Each person tested, teramsuk learners, certainly has a sense of anxiety even though the amount varies. If there are learners who have always had high anxiety levels when tested, the measurement results tend to underestimate of the actual ability. In measuring, the teacher can make systematic errors. This error can occur when the scoring, there are teachers who "generous" and there are teachers who "expensive" in scoring. Cheap and expensive if it applies to all learners, there will be a systematic error. Sebalikya, when it applies only to certain students, there will be a bias in the measurement. So complex measurement problems that required measurement theory. Currently there are two theories of measurement used widely, namely classical test theory and modern theory. Classical test theory assumes that a person's scores obtained from the results of a measurement can be broken down into actual scores and scores of errors. Other assumptions are that there is no relationship between the actual scores with scores of errors. The second basic assumption of this, further developed formulas or formulas to determine the validity index (validity) and the index of reliability (reliability).
There are some weaknesses classical test theory, and the
most prominent is the statistical dependence on the characteristics of the item being measured. Thus, the amount of grain statistics vary from one group to another group. As a result, it is difficult to compare the ability of one group to other groups, especially among individuals. This weakness has long been recognized, that since the development of measuring instruments used in the field of natural sciences or technology. Measuring tool used in this field does not depend on the object being measured, because its characteristics do not change over the same measured object. It is easy to understand because the measured object or objects are dead. In contrast to the object in the field of education, namely humans. The human condition as happy and difficult conditions, always changing from time to time, so that the measurement results obtained do not necessarily indicate the actual individual characteristics. Therefore, developed the theory of measurement that can overcome the weakness of the classical theory. Classical theory that developed at the moment -which is called the theory of modernization using some basic assumptions. The underlying assumption is someone the opportunity to answer correctly an item not ditentuka by the opportunity to answer another item, known as independent assumptions. The modern theory seeks to develop an analysis that produces estimates of a person's ability without being influenced by the measuring instruments used. Likewise, the statistics point endeavored not depend on individual characteristics measured. Based on these characteristics, modern test theory developed by many experts in the measurements of this world.