Home Education Handbook of Test Development
Ideally, test takers should have people like them represented in the test, if the items include people. For subject-matter tests, your ability to represent diversity will be determined by the subject. For example, you will have a greater opportunity to include a variety of people in a literature test than in a physics test. For skills tests, such as reading comprehension, there are fewer subject-matter constraints and you will be able to include a greater variety of people.
Avoid items that reinforce stereotypes. Items showing traditional behaviors (e.g., a woman caring for a child) are acceptable, but they must be balanced by items showing members of the group in nontraditional roles. If you show only traditional behaviors for a group, you are reinforcing stereotypes.
Because validity and fairness are so closely intertwined, reviews for the extent to which tests meet their specifications and the extent to which construct-irrelevant variance has been excluded serve to enhance both validity and fairness. In addition, review completed tests to ensure that the diversity goals have been met and that any potentially stereotyped depictions of people have been balanced by nontraditional representations. For adaptive tests, make the check for appropriate diversity and balance in the item pool. Also check a few sample forms as they would be administered to high-, medium- and low-scoring test takers.
Challenge items that are out of compliance with the guidelines, even though the items had previously passed fairness reviews. Guidelines change over time. Furthermore, some topics that were acceptable when the items were written may become controversial or elicit very negative emotions because of later events.
|< Prev||CONTENTS||Next >|