6. Assessment must be fair and equitable.
We live in a multicultural society with laws that promise equal rights to all. Our school communities must work to ensure that all students, as different as they are in cultural, ethnic, religious, linguistic, and economic background, receive a fair and equitable education. Assessment plays an important part in ensuring fairness and equity, first, because it is intimately related to curriculum, instruction, and learning, and second, because assessment provides a seemingly impartial way of determining who should and who should not be given access to educational institutions and resources. To be fair, then, assessment must be as free as possible of biases based on ethnic group, gender, nationality, religion, socioeconomic condition, sexual orientation, or disability. Furthermore, assessment must help us to confront biases that exist in schooling.
In the past, standardized tests have been viewed as a means to avoid the cultural and personal biases of teachers’ judgments. However, just as it is impossible to eliminate bias from teachers, it is also impossible to produce an unbiased test of reading or writing. Language itself involves social conventions that differ from culture to culture. Furthermore, words have different shades of meaning for different cultures, and the variation in life experiences across culturally, economically, and geographically different situations can be quite extreme. Consequently, students differ enormously in the interpretations they give to the texts they read, the topics they feel comfortable writing about, and the ways they respond to different forms of assessment. The curriculum-distorting effects of high-stakes testing are also distributed unevenly across subgroups of the population. In the United States, urban schools with significant numbers of students living in poverty are more subject to the curriculum-narrowing pressures of high-stakes testing than are more affluent suburban schools.
The inevitability of bias notwithstanding, when tests must be used, as many biases as possible should be controlled. Whenever possible, assessment should be accomplished in a language that will not interfere with the individual’s performance. Assessment practices should not devalue cultural differences in dialect. Students have the right to learn the language of the dominant culture because it is the language of power. However, students should not be penalized in assessments for using their home language where the privileged dialect is not specifically required. Assessment must also take into consideration the differences between basic and academic language and the length of time students need to become skilled at each.
Biases routinely occur in assessments and in the curricula they represent. For example, all students should study and be assessed on literature from and knowledge of cultures other than their own. Failure to do so introduces a cultural bias. However, there are other biases that regularly occur as a result of assessments. Students who are initially less successful than others in literacy acquisition often find that their curriculum shrinks to one that is less engaging and less mind-expanding. This form of bias is often also associated with economic differences across schools, and it perpetuates those differences by reducing the breadth and complexity of the literacy students acquire. Assessment that allows for critical inquiry into the curriculum is an important antidote to such common but avoidable inequalities and also serves to make institutional biases clear and public.
Most biases are part of the perspective we bring from our cultural backgrounds, so we tend not to notice them in ourselves. We must strive to have the testing industry, policymakers, administrators, and teachers—all those charged with creating and interpreting tests—reflect and respect the diversity of our society. At the same time, it is particularly important that multiple perspectives be brought to bear on assessment issues (see standard 8). One way to take test bias seriously would be to ensure strong and varied representation of culturally, ethnically, linguistically, and economically diverse groups in the construction of public tests. In this way, test biases should become apparent and, once recognized, be easier to reduce. A second important way to address bias is to make tests available for public examination after they have been given. A third way to offset bias is to ensure that no single assessment is used to make important educational decisions (see standard 8).
Inequities in schooling can also be compounded through inappropriate assessment. For example, assessment practices, both large scale and centered in the classroom, often lead to students being placed in different instructional settings or programs with the intention of producing a better match between student and curriculum. This leads to a significant equity issue. On the one hand, a better instructional match is possible, but on the other, different and perhaps lowered expectations on the parts of both teachers and students themselves may result. Once students are assigned to systematically different curricula, uneven access to subsequent experiences and jobs becomes not just possible, but probable.
Other uses of assessments can also produce inequities. For example, external pressures regarding the use of tests often differ across school settings within individual districts or specific regions. This is particularly common in large cities. Similarly, a common practice in newspapers in some areas is to report the average test scores of students by district, school, or even classroom. Because individuals and businesses are reluctant to move into areas where schools have low scores on tests, tax bases and economic resources erode in these neighborhoods with the result that economically stressed school systems become more so. Pressure on teachers also increases, which creates greater teacher attrition and leaves high-needs schools with a less experienced teaching force.
When assessing, we must be sure to attend to the relevant competencies. For example, provisions should be made to ensure that second-language learners are assessed in ways that permit them to show what they know and can do, with consideration for the time it takes to develop both basic and academic language. For students classified as reading disabled, the situation is less clear. In some U.S. states it is considered appropriate for these students to have their reading assessments read aloud to them. This practice may seem fair, but it makes productive inquiry impossible because the assessment no longer represents the construct “reading.”
We must also remember that, although assessment plays an important role in ensuring fairness and equity, the goal of equity cannot be laid solely at the feet of assessment. No assessment practice can shore up the differences in educational experience that arise from the obviously unequal conditions of extreme poverty and wealth.