Philosophy and measurement of school achievement

Journal of Instructional Psychology, Sept, 2003 by Marlow Ediger

Commercial Tests Used to Measure Pupil Achievement

Commercial companies develop standardized achievement tests, also called norm referenced tests (NRT). These are used rather frequently to measure pupil achievement. There are different elements that go into the development of standardized achievement tests since the conditions for test taking are to be the same for all pupils. The following factors are then to be the same:

* time limits enforced for taking each test.

* subject matter tested on for each set of pupils taking the designated test.

* directions given to pupils for test taking.

Professional writers write the test items for a specific curriculum area or areas. These test items then are responded to by pupils in a pilot study. Feedback from the pilot study provides knowledge, to the test writers, of the quality of each test item. Computerized scoring using the same key further standardizes the test in that conditions are the same for all involved in taking the test. When writers of the test items see the computerized printout from pilot study results, they notice which test item was too easy and all responded correctly to that test item. Or the other extreme whereby all responded incorrectly to a test item. Where all responded correctly or incorrectly to the same test item makes it either too easy or too difficult. Then too, pupils results are not spread out adequately. Generally, the lowest and the highest 5th percentile scores in a pilot study are eliminated to take care of extreme scores. When an adequate number of pilot studies have been run, then a spread of scores will result from the first to the 99th percentile. Standard tests are designed to spread out pupils' scores from high to low. Each academic or curriculum area will have data on pupils' scores and their equivalency in the Manual of the standardized test.

Quality validity and reliability will be discussed in the manual and how the were determined. The test writers are quite far removed from the local classroom in ascertaining which test items should be written and included on the test. The test will be administrated in various parts of the United States to pupils from diverse levels of achievement. Local pupils who have taken the standardized test will have their scores compared with those in the Manual (who were in the final pilot study) to notice the related percentile level.

Criterion Referenced Tests

To take care of the predetermined methods of spreading scores out from high to low, as did advocates of standardized testing, criterion referenced testing (CRT) was invented and brought into the measurement movement. The CRTs have objectives available which teachers may use as guidelines for instruction. CRTs are pilot tested to take out vaguely written test items. Clearly written, meaningful test items should be in evidence. As is true of standardized tests, CRTs also tend to use multiple choice test items Together with the stem, each distractor of the four in a multiple choice test item needs toe plausible. The four responses should be of similar length. No, clues should be given as to which is the correct response.