Sources of test bias are discussed and steps to prevent or reduce bias in tests are listed. Test bias can occur because of the way test materials are written, the conditions of administration, and the interpretations given the results. Steps to prevent or reduce bias arising in the test development process include: (1) using heterogeneous sets of…

Asserts that publishers of achievement tests are, for the most part, not in a position to obtain on their own any decent evidence about the consequences of uses made of their tests. Reasons why this is so are discussed, and what publishers can be expected to do is outlined. (SLD)

Potential benefits of using item response theory in test construction are evaluated using the experience and evidence accumulated during nine years of using a three-parameter model in the development of major achievement batteries. Topics addressed include error of measurement, test equating, item bias, and item difficulty. (TJH)

In determining how reliable is reliable enough and how much error can be tolerated in criterion-referenced testing, the following relationships hold: (1) the more specific an objective is, the fewer the items required to reliably measure it; (2) the more specific the objectives are, the more objectives required to cover a given span of the…

Discussions of "test curriculum overlap" that focus on the term "mismatch" tend to be based on several untenable premises. This paper addresses the issue of the degree to which standardized tests should reflect the specific content of a given school curriculum with regard to three points: (1) The idea of matching the content of…

Improved communication and cooperation between test publishers and school systems is essential for solving major test problems. Test-curriculum matching, test equating, and inappropriate use of grade equivalent scores are major problem areas. The degree to which a standardized test yields information relevant to what the system is trying to teach…

As the new version of the "Standards for Educational and Psychological Testing" is being developed, it is apparent that putting together a set of standards for test publishing involves many difficulties. Although the basic intent of almost all parties involved is similar, there are many potential areas of disagreement among parties to…

Potential benefits of using item response theory in test construction are evaluated, based on the experience and evidence accumulated during 9 years of using a three-parameter model in the construction of major achievement batteries. Specific benefits covered include obtaining sample-free item calibrations and item-free person measurement,…

This paper describes the development of a criterion-referenced test. The Prescriptive Mathematics Inventory (PMI) was developed to measure 400 stated behavioral objectives. The test consists of three overlapping levels with the objectives chosen to cover 90 to 95 per cent of the mathematics curriculum nominally taught in grades 4 through 8. Each…

During the past few years the problem of bias in testing has become an increasingly important issue. In most research, bias refers to the fair use of tests and has thus been defined in terms of an outside criterion measure of the performance being predicted by the test. Recently however, there has been growing interest in assessing bias when such…

The use of item banks and item response theory has resulted in new ways to misinterpret and misuse tests through customized, yet standardized, achievement test batteries. The new test batteries create the possibility of serious misunderstandings based on the idea that any subset of items from the pool with a proper range of difficulties will…

The passage dependence of an item refers to the degree to which it is necessary to read the associated passage in order to be able to answer that item correctly. This paper studies the value of obtaining data about passage dependence in the construction of reading comprehension achievement tests. Passage dependence indices were examined in several…

This paper is concerned with the accusations made by such groups as the Association of Black Psychologists in their call for a moratorium on testing, that standardized tests are biased. A biased test measures one trait in one group of people but a different trait in a second group. Evidence about the amount of bias in tests is thin. Bias must be…

A description of two proposals for alleviating the racial and ethnic bias in tests of achievement used in schools is presented. One of them entails adding steps to the construction procedures used in building norm referenced achievement tests; the second entails using criterion-referenced achievement tests rather than standardized tests for…

Basic skills tests in which measurement of growth is defined in the context of national norms are discussed. An integral part of the construction of norms is the production of a trait or ability score which is normed. Test publishers define a multidimensional trait by including items which measure it. The trait as a construct is a distinct…