| Home |
 |
Chapter 13 |
|
Theory into Practice |
|
Ch. 13, p. 479
Making Assessments Fair
Although fairness in assessment is something everyone believes in, defining fairness in assessment is not straightforward. Indeed, the latest edition of the Standards for Educational and Psychological Testing gives four definitions and acknowledges that many more are in the literature (AERA/APA/NCME, 1999). For our purposes, we might define fairness as being honest, impartial, and free from discrimination.
Besides being ethical, fairness makes good instructional sense. Fair testing encourages students to spend more effort on learning because they will come to see that success depends only on what they know and can do.
Fairness in assessment arises from good practice in four phases of testing: writing, administering, scoring, and interpreting assessments. Practices that lead to fairness in these areas are considered separately below.
Writing assessments. Base assessments on course objectives. Students expect a test to cover what they have been learning. They also have a right to a test that neither “tricks” them into wrong answers nor rewards them if they can get a high score through guessing or bluffing.
Cover the full range of thinking skills and processes. Assuming instruction has included higher order thinking, so should an assessment based on that instruction prompt students to use the material intellectually, not merely repeat memorized knowledge. Further, if a teacher’s tests cover only memorization, the students will emphasize only memorization of facts in their preparation.
Cover course content proportionally to coverage in instruction. The content areas on the test should be representative of what students have studied. The best guide to appropriate proportions is the relative amounts of instructional time spent on those topics.
Test what is important for students to know and be able to do rather than isolated trivia. The best guide to the content most appropriate for the test is to cover what is important for students to come away with from the course. When writing a test, ask yourself whether each task is what other teachers would agree is important when teaching that course. Better, ask a colleague to review your draft test.
Avoid contexts and expressions that are more familiar and/or intriguing to some students than to others. One challenge in writing tests is to make sure none of your students are advantaged or disadvantaged because of their different backgrounds. For example, music, sports, or celebrity-related examples might be appealing to some students but not others. Language or topics should not be used if they are more well known or interesting to some students than to others. If that proves impossible, then at least make sure the items that favor some students are balanced with others that favor the rest.
Giving assessments. Make sure students have had equal opportunities to learn the material on the assessment. Whether or not students have learned as much as they can, at least they should have had equal chances to do so. If some students are given extra time or materials that are withheld from others, the others likely will not feel they have been treated fairly.
Announce assessments in plenty of time for students to prepare for them. Since students’ learning styles differ, some will keep up to date with their studying and others will prefer to put in extra effort when it is most needed. Surprise assessments reward the former and punish the latter. But these styles are not part of the material to be learned. Not only is it more fair to announce assessments in advance, it also serves as a motivator for students to study.
Make sure students are familiar with the formats they will be using to respond. If some students are not comfortable with the types of questions on an assessment, they will not have an equal chance to show what they can do. If that might be the case, some practice with the format beforehand is recommended to help them succeed.
Give plenty of time. Most tests in education do not cover content that will eventually be used under time pressure. Thus, most assessments should reward quality instead of speed. Only by allowing enough time so virtually all students have an opportunity to answer every question will the effects of speed be eliminated as a barrier to performance.
Scoring assessments. Make sure the rubric used to score responses awards full credit to an answer that is responsive to the question asked as opposed to requiring more information than requested for full credit. If the question does not prompt the knowledgeable student to write an answer that receives full credit, then it should be changed. It is unfair to reward some students for doing more than has been requested in the item; not all students will understand the real (and hidden) directions since they have not been told.
Interpreting assessments. Base grades on summative, end-of-unit assessments rather than formative assessments that are used to make decisions about learning as it is progressing. The latter are intended as diagnostic and to be used to help accomplish learning. Since grades certify attainment, they should be determined based on assessments made after learning has taken place.
Base grades on several assessment formats. Since students differ in their preferred assessment formats, some are advantaged by selected-response tests, others by essay tests, others by performance assessments, and still others by papers and projects.
Base grades on several assessments over time. As with assessment formats, grades should also depend on multiple assessments taken at different times.
Make sure factors that could have resulted in atypical performance for a student are used to minimize the importance of the student’s score on that assessment. If it is known that a student has not done her or his best, then basing a grade or other important decision on that assessment is not only unfair; it is inaccurate.
|