Book Description
In this book, the authors provide a cogent review of statistical and interpretive procedures that, in combination, can be used to reduce the likelihood that tests contain items that favor members of one gender, age, racial, or ethnic group over equally able members of another group, for reasons that are unrelated to the objectives and purposes of measurement. Such test items are said to be biased against the equally able members of the group that is not favored. The methods described and illustrated in this book have the potential to reducing the incidence of tests that are, in their construction, biased against members of one or more groups. These methods have the potential of controlling an important source of invalidity when test results are interpreted.