The Tracking Decision
Your department is deciding whether to keep or dismantle its two-track Algebra I sequence (honors and regular). Data from the past three years: the honors track is 71% white and Asian, the regular track is 68% Black and Latino, in a district that is 44% Black and Latino. Teacher recommendations drive placement, and the recommendation form asks for 'work ethic' and 'maturity' alongside math grades. A parent group has raised concerns. The department chair asks you to 'look at the data and give a recommendation.' You teach one section of each track.
1. Analyze the placement data and the recommendation form for evidence of bias, being precise about what the data do and do not show.
Read the worked answer
What the data show:
- A large demographic skew between tracks (71% white/Asian in honors vs. 68% Black/Latino in regular) in a district that is 44% Black/Latino. The skew is large enough that it is unlikely to be explained by chance or by prior achievement alone - but the data as given do not include prior achievement, so I cannot yet separate 'placement reflects preparation' from 'placement reflects bias.'
- The recommendation form is the stronger signal. Asking teachers to rate 'work ethic' and 'maturity' alongside math grades introduces subjective criteria that are not math achievement and are well-documented sites of bias. Even well-intentioned raters show bias on such scales. The form mixes an objective signal (grades) with subjective ones, and if the subjective ones carry weight, the placement is not purely achievement-based.
What the data do not show:
- They do not prove individual teachers are biased.
- They do not prove that dismantling tracking would improve outcomes - that's a separate causal claim.
- They do not control for prior math performance, which is the obvious confound.
Honest conclusion: The demographic skew is a red flag requiring investigation, and the recommendation form is a specific, fixable mechanism that plausibly contributes. I should not overclaim causation, but I also shouldn't dismiss the pattern.
2. Propose a concrete recommendation to the department, with at least one change you can implement immediately and one that requires a longer timeline.
Read the worked answer
I'd recommend not dismantling the tracks immediately (that's a multi-year change with real costs and needs evidence) but fixing the placement mechanism now, because that's where the bias risk is concentrated and where I can act.
Immediate (this placement cycle):
- Remove 'work ethic' and 'maturity' from the recommendation form. Replace with math-specific, observable criteria: performance on specific problem types, a common placement assessment, and a student/parent opt-in pathway. This directly addresses the subjective-bias mechanism.
- Add a universal screening step: every student takes the placement assessment; teacher recommendation becomes one input, not the gate. This removes the 'teacher must nominate' bottleneck that can exclude students nobody thought to recommend.
- Audit last year's placements: for each student placed in regular who scored above the honors threshold on the common assessment, ask why. This is a concrete, small data pull.
Longer timeline (1–2 years):
- Pilot an open-enrollment honors section with support structures (summer bridge, in-class scaffolding) and compare outcomes to the current honors section. This generates local evidence on whether dismantling helps, rather than importing a claim from elsewhere.
- Track placement data by demographic each cycle and report it to the department, so the pattern stays visible.
Rationale for sequencing: The form change is low-cost, high-leverage, and defensible regardless of how the tracking debate resolves. The pilot answers the causal question the current data can't. Doing both protects against the failure mode of either doing nothing or making a sweeping change on thin evidence.
3. A colleague says, 'Our honors track is fine - the demographics just reflect who does the work.' How do you respond in a way that keeps the conversation productive?
Read the worked answer
I'd avoid a moral framing ('that's biased') and instead make the claim testable - which is also the most respectful move, because it treats my colleague's view as a hypothesis to check rather than a prejudice to correct.
What I'd say: 'That's a real possibility, and it's testable. If the skew is about who does the work, then students placed in regular who do the work should be rare. Let's check: of last year's regular-track students, how many scored above the honors threshold on a common assessment or earned an A in Algebra I? If it's near zero, you're right. If it's a meaningful number, we have students who did the work and were still placed in regular - and that's worth understanding.'
Why this works:
- It accepts the colleague's premise as a falsifiable claim rather than dismissing it.
- It shifts from opinion to a specific data pull that either of us can run.
- It keeps the focus on students who might be misplaced rather than on teachers who might be biased - which is where the conversation can actually move.
What I'd avoid: citing national research at them (feels like a lecture), implying they're racist (shuts down), or conceding the point to keep peace (abandons the students).
Follow-through: I'd offer to run the audit with them, so the result is shared ownership rather than my finding imposed on them. If the audit shows near-zero misplaced students, I've learned something and should say so. If it shows a meaningful number, we have a shared problem to solve - and the form change becomes an obvious next step.