UC Oceanview — Debrief & Reflection

Debrief

Can Your Preferred Option Justify Its Cost?

You've now explored six removal criteria — five that use identity-neutral standards and one that uses none at all. Every criterion produced a different pattern of adverse impact. Every criterion harmed someone. And the lottery — the only option with no criterion — was the only one that harmed no one disproportionately.

Random selection is not a serious policy proposal. No university would remove 2,800 admits by coin flip. But it establishes the benchmark: this is what zero adverse impact looks like. Every deviation from that benchmark is a choice — a choice to impose disproportionate harm on specific communities in pursuit of some other value. The question is not whether that choice is defensible. The question is whether the decision-maker knows they are making it.

Five Questions

What Did You Learn?

1 The lottery question If a criterion-free process produces zero adverse impact, then the adverse impact you observed under Options 1a through 4 is not an inevitable feature of reducing the class by 2,800. It is a feature of the criterion you chose. What does your preferred criterion value — and is that value worth the demographic cost it imposes?

2 The denominator question Option 1a imposes +14.8 pp of adverse impact on Hispanic admits. Option 1b — the same construct, measured contextually — imposes +7.7 pp. Roughly half. The reduction didn't require abandoning academic rigor as a criterion. It required measuring rigor in context. How much adverse impact is your organization willing to inflict because it hasn't changed its denominator?

3 The convergence question Options 1a and 2 both harm Black and Hispanic admits — through entirely different mechanisms (school funding for AP/IB access vs. grading standards and course rigor for GPA) but with the same directional result. Why do most "neutral" criteria point in the same direction? Because structural inequality in K–12 education runs in one direction. The criteria are neutral. The structure is not.

4 The negative action question Option 4 harms Asian admits — and still clears the four-fifths rule, at a ratio of 0.82 against a 0.80 threshold. The bias embedded in holistic review is real but legally invisible. Eliminating affirmative action did not eliminate this bias. It has nothing to do with affirmative action. It is negative action: the presence of a ceiling, not the absence of a floor.

5 The transfer question Where in your organization do you see the same tradeoff — identity-neutral criteria producing disparate impact through structural correlation? What is your Option 1a? Your Option 3? And what would the contextual-rigor equivalent look like — measuring effort relative to opportunity rather than absolute output?

How the Numbers Were Computed

Reading the Measures

Adverse impact (pp) = a group’s share of those removed minus its share of the 8,000-admit pool. The label carries the negative; the sign carries direction and size. Positive means the group was over-removed, the criterion taking more of it than its share of the pool — under Option 1a, Hispanic admits run +14.8 pp. Negative means under-removed — Asian admits run −15.3 pp under that same option. Zero is the lottery.

The four-fifths screen divides each group’s retention rate by the highest group’s rate and flags any ratio below 0.80 (EEOC Uniform Guidelines, 29 C.F.R. § 1607). It is a legal threshold, not a finding of fairness. Option 4 is the case worth sitting with: Asian admits absorb +4.5 pp and clear the screen at 0.82.

Total disparity sums the absolute adverse impact of the four largest groups — White, Asian, Hispanic or Latino, and Black or African American admits, 92.1% of the pool. Because it sums magnitudes, it carries no sign: it runs from zero upward, and higher is worse. The four smaller groups are excluded because at 45 to 316 admits their percentage-point values swing on a few students. That is a measurement choice with a cost, and it is worth asking what a measure that suppresses the smallest groups is doing when the American Indian or Alaska Native flag under Option 4 is the one the four-fifths screen catches.

Cross-Domain Applications

Where Else Does This Happen?

The mechanisms surfaced by UC Oceanview operate wherever organizations use standardized criteria to sort, evaluate, or select people. The specificity of the admissions context makes the dynamics visible. The framework generalizes.

UC Oceanview Option	Organizational Analog	Key Question
Option 1a Absolute AP/IB	Requiring specific certifications, credentials, or training completions for promotion — excludes employees without access to development programs	Does your organization measure absolute credentials or credentials relative to opportunity?
Option 1b Contextual Rigor	Evaluating employee performance relative to role difficulty, team resources, or market conditions rather than absolute output metrics	What would the denominator be? How would you measure "opportunity to perform"?
Option 2 Unweighted GPA Floor	Using a single performance-metric cutoff as the sole criterion for layoffs, hiring, or advancement	What does the metric measure — ability, preparation, or access to preparation?
Option 3 Geographic Caps	Regional hiring quotas, location-based talent allocation, or office-level headcount caps applied to a diverse workforce	Which locations bear the cost of a geographic constraint? Is that cost correlated with the composition of those locations?
Option 4 Holistic Review	Manager discretion in talent reviews — subjective evaluations that pass the four-fifths rule but embed implicit bias in who is rated as "high potential"	What structural advantages flow to the people managers already know?
Option 5 Lottery	Random assignment to desirable roles, projects, or development opportunities as an equity benchmark	Under what conditions is randomness more equitable than judgment?

"The criteria are neutral. The structure is not."

Identity-neutral ≠ identity-irrelevant. How you operationalize a criterion determines who it harms. And the choice of denominator is itself an equity intervention.

← Back to the lottery ← Back to all options