ACHIEVEMENT TEST, a set of questions or problems designed to determine how much an individual knows about some subject area. Broadly speaking, the term would cover the examinations school teachers prepare and use as a basis for assigning grades, the bar examinations lawyers must pass before they are permitted to practice, the oral examination Ph.D. candidates must pass, and many other testing situations. Ordinarily, however, the term “achievement test” is limited to standardized, published instruments designed and constructed by specialists in measurement techniques. Such tests are given in arithmetic, English grammar, biology, history, and many other school subjects.
The development of an achievement test involves several steps. The fîrst is to delineate the area the test is intended to cover. The usual practice is to assemble a committee of outstanding teachers of the subject. They outline the topics about which cjuestions are to be written and determine what proportion of the questions should be focused on each topic.
The second step is to write the questions or problems. They should be written in such a way that only accurate knowledge and an appropriate thinking process will produce correct answers. It has proved possible to involve many kinds of reasoning processes in objective test items such as true-false and multiple-choice questions. Thus achievement tests are not limited to measuring simple factual knowledge but can assess such complex mental skills as judgment, application of principles, and organization of ideas.
The third step is statistical. The test under construction is administered to a trial group similar to the people for which it is intended, and an item analysis is made. All the answers given by the members of the trial group are tabulated. From these tabulations indices are obtained showing how difficult each item is for the group in question and how well the items discriminate between good and poor students of the subject. Various kinds of errors and ambiguities in the items also become apparent at this stage. Often it is necessary to rewrite the test and repeat the item analysis. This trial stage makes it possible to determine the overatt reliability or accuracy with which the test is measuring the characteristic it is designed to measure.
The fourth step is to administer the revised test to a large group chosen in such a way as to be representative of the people on whom the test is to be used—for example, seventh graders in the public schools, or entering college freshmen. From the scores of this group, norms of some kind are developed—age norms, grade norms, percentile scores, Standard scores, or some combination of these. How well an individual performs on the test can be judged by comparing his score with one or more of the norms.
In evaluating an achievement test, the primary consideration is content validity. This depends primarily on the competence of the persons who made the initial decisions about the makeup of the test and on the skill with which the educational objeetives were incorporated in the test items.
Achievement tests are used for many purposes such as educational research, educational planning, placement of individuals in special elasses and seetions, seleetion of students for college admission and for seholarships and awards, and as a basis for a person’s own analysis of his potentialities and handicaps. In short, they are devices to facilitate educational decisions about individuals and by individuals.