Lessons for Higher Ed from K-12 Standardized Assessment
Dr. Catherine Taylor, professor emerita from the University of Washington, spent a career working in and studying K12 state assessment. She observed everything from implementation of No Child Left Behind in the 1990s to computer adaptive testing in recent years. She has seen standardized exams scores improve across sub-populations and revert back; not all K12 changes are “forward” progress toward equity. So what lessons from K12 assessment can we learn from for higher education?
Constructed response items in standardized tests. Girls and underrepresented populations are systematically suppressed by standardized multiple choice tests; however, incorporating constructed response items lessens the gaps. Constructed response items are based in reader response theory allowing an individual students’ background and experiences to interpret the question and provide an appropriate response, especially when items ask students to interpret text, data, or results from scientific investigations. Higher education should be paying attention to this not only to consider expansion to incorporate more constructed response exams but because of downstream impacts on students’ confidence and sense of belonging in academic programs that have strong math or science roots. The disproportionate numbers of women and underrepresented students in these fields still indicate a need for cultural shifts in these fields.
Bias and sensitivity reviews. K12 in Washington state utilizes stakeholders who are not just token members of populations, but representatives from these groups that have an expertise or experience with the population or are active in the culture. These groups are brought together to read test items to help identify bias or sensitivity issues that could disadvantage a certain population. As Dr. Taylor says, “We need to listen because stakeholders will teach us things.”
Stakeholder engagement. Dr. Taylor would like to see more research partnerships between state governments and academics. Academics can provide assistance and analysis power to help evaluate state testing procedures and analyze results. We need to move politics aside and work together to improve student outcomes. An edited volume sponsored by the National Council on Measurement in Education, Culturally Responsive Assessment in Classroom and Largescale Contexts, is to be published by Routledge in late 2024 or early 2025. The volume includes chapters by prominent assessment specialists on current research, theory, practice, and policy relevant to these bias issues. In particular, all of the authors in the volume strongly urge the involvement of minoritized groups in all phases of assessment development: from determining educational goals to item writing to item content and bias reviews to developing scoring algorithms to score reporting.