Commentary: Social Responsibility in College Admissions Requires a Reimagining of Standardized Testing

As college admissions becomes more competitive in the United States and globally, with applicants competing for limited seats, many programs are transitioning away from standardized testing as an application requirement, part due to concern that can perpetuate inequities among increasingly diverse student population. In this article, I argue we only address by reimagining ground up. Following a summary of recent debate around at University California (UC), discuss how my perspective aligns Koljatic et al. (2021), who encourage industry accept social responsibility. Building on themes focus article other publications, then propose that, contribute educational equity, must work toward is transparent openly accessible than ever before. Over past 12 months, longstanding alliance between UC its preferred admission test, SAT, has itself been subject high-stakes examination. spring 2020, roughly two months into global pandemic following scandals (Jaschik, 2019) litigation (Public Counsel, 2020) previous year, Board Regents voted unanimously via Zoom call suspend SAT/ACT requirement (UC Office President, 2020a, 2020b). Their resolution left open possibility finding or creating alternative test 2025. However, event was not feasible such compressed timeframe, would eliminate requirements entirely. This addressed growing concerns disparate impact underrepresented students (e.g., Geiser, 2017), but it also nearly orthogonal recommendations Academic Senate (Douglass, 2020). The had organized Standardized Testing Task Force, which concluded through comprehensive study benefits testing, including groups, outweighed costs, even selection purposes (Systemwide Senate, UC, fall convened what they called feasibility group, member, evaluate options replacement test. We determined although existing Smarter Balanced system adopted California) could support lower-stakes components process like determining eligibility, be develop new effectively 2025, any recommended 2021). present background information highlight one consequential transitions terms number involved (roughly 118,000 admitted 2020 over 170,000 applicants; Infocenter, case demonstrates range views regarding our next steps should be, arguments (a) better tests Cai, 2020), (b) repurposed end-of-year Kurlaender, Reber, & Rothstein, (c) no (specifically, high-stakes, norm-referenced testing; e.g., 2017; most committee review takes view step further, confirming lack confidence equitable Thus, final examination exams, do current approaches meet expectations, approaching and, least, door seems closed solutions, long stakes high. their (2021) briefly controversy surrounding reference national exams Chile, outline responsibility respect commercial tests. agree general message extensions validity Sireci, itself, measurement community broadly, achieving own intended purposes, requires proactive design implementation order reduction inequities. If contribution within limits purpose, revisit may involve redefining constructs (Randall, 2021) reconsidering necessary extent standardization (Sireci, How apply specific context testing? By way analogy, refer shift corporate deniability exhibited Nike, Inc. 1990s. multinational footwear apparel company faced intense backlash when refused abusive working conditions overseas manufacturing. Nike argued manufacturing subcontract thus outside control, eventually impacted sales accepted took improve (see Spar La Mure, 2003; Burns, 2002; Nisen, 2013; referenced al., outlines similarly “develop consciousness level about products affect customer base” “integrate improvement gap business model” (p. 3). At risk pushing analogy beyond reasonable limits, dig deeper comparison industries, acknowledging first issues each differ complexity. Improving worker reasonably sphere influence contracting work. assume others done so, solution clear. On hand, reducing score gaps much less tractable, research showing variety factors differential academic achievement (Carter, 2018). say themselves those factors, right back observe. Instead, solutions will require partnerships well community, K12 education public programs. Along these lines, recognize essentially subcontractor state university systems, building instruments according specification. have direct experience development, yet integrated model because, product specification, because intentionally excluded stated principles, example, include “fairness follows uniform comparisons across whole applicant pool” 2009, p. 9), words, fairness norm-referencing made possible standardization. So, whereas appears initially chosen convenience personal gain, assumed neutral position, part, were expected required so. To clear, does claim solely responsible performance gaps. They acknowledge start “college cause adverse tests” 2; emphasis original), promote multiple stakeholders. seem suggest currently taking defensive position benefit, apparently did, need true. Regardless complexity problem, regardless motives blame shared different stakeholders, take active role solution. Demand primarily single-use, limited-access, changing. balance all treated same” everyone individuals” 4) latter, signaling highly standardized, normative interpretations criterion-referenced ones suggests some adapt. below then, considering magnitude issue, go further. recommend related reorienting testing. First, attend closely content, ensure knowledge skills needed college, opposed broad curricular coverage, specified standards. recommendation trend states using place grade eleven federally mandated (Lorié, Second, measure account opportunity learn interpreting results. response recommendation, case, where alignment curriculum standards identified key principle addition fairness, reliability, predictive power, utility “that justifies monetary costs,” Admissions Relations Schools articulates “an levels mastery content UC-approved high-school preparatory coursework provide students, parents, educators, enabling them identify strengths weaknesses” 18). Emphasis placed standards, rationale broadly aligned know able produce might covered strictly Criterion-referenced, curriculum-based course planning remediation. Aligning drawbacks. notes (when serving struggle well, quality instruction depend heavily socioeconomic status). Force mentioned above discussed limitations issue construct being tested, versus aptitude Zwick, 2004). article's second measuring learn, often holistic wherein professionals already attempt evaluation transcripts courses available applicant's high school (the incorporates Eligibility Local Context; 2009). potentially processes these, commercialized measure, developed maintained external program, preferable measures managed internally. While worth consideration, dramatic changes challenges face. below, shifting mission organization reimagine criticized reasons, financial cost takers, reliance narrow administrations, preparation, bias individual items, marginal power chapters 2004 discussion). Tests viewed inauthentic inaccessible populations (Arbuthnot, 2020; Randall, 2021; Trumbull Nelson-Barber, 2019). Admission forms applied, place, contexts detract learning differentially limit opportunities (Geiser, 2017). Though safeguards typically aim protect against misuse, still seen outweighing costs. Critics proponents disagree whether problematic negligible, threaten effective use (Zwick, there consensus fact, having positive hope will. After decades States, persist. Sireci (2020) recommends free ourselves constraints pursuit flexible accurate takers. Extending Sireci's (2020;2021) bold vision, drawing culturally responsive pedagogy (Gay, 2018) (Jhangiani Biswas-Diener, path impacts generally, integrate crowd-sourced development improvement, fully mean item banks freely online. Of pursue, open-access potential impact, removes veil secrecy encourages us alone (Rohrer Pashler, 2010). occasional release sample items insufficient. constrained technology, goal build so deep capture entire domain universe task, practice later summative without security concerns. Administration, scoring, scaling procedures designed needed, depending application. organically dynamically quality. Not promise efficient selected but, training, empowers directly assessing sensitivity, accessibility. Furthermore, population, target access facilitate rapid representative engaging historically marginalized groups. involvement replace diminish expert review. It supplement it. Finally, transparency results, deidentified data readily analysis level, rely developer intermediary evidence Similar too empower informed consumers information. A given extreme form, derive benefits. open-access, engagement, results goals toward. expect follow. These buy-in users consumers; expansive, diverse, authentic, content; decrease preparation corresponding increase groups continuous full closing, return another analogy. recently released shoe put taken off hands-free, extending lineup (Newcomb, innovation regarded major forward, speak, inclusive individualized (Patrick Hollenbeck, raised accessibility availability (Weaver, compare ways, example highlights very least consideration tests, large-scale assessments, design, until moment administration. Integration assessment systems significantly richer decisions Kurlaender innovations (The Gordon Commission Future Assessment Education, 2013). Yet, despite advances, largely controlled conditions, inclusively shoes rented worn certain occasions under supervision. Our vision distribute ownership itself. interest space, leave discussion obstacles many, future paper. outlined ambitious, lead hardly resemble now familiar with. But challenge face existential one, emphasized demands disruptive innovation. colleges universities, sent clear unacceptable, proposed Systemwide appear far enough persistent system. Let explore