Two Graders, One Rubric, One Appeal

The pass-rate gap (38% versus 8%) is far too large to be explained by student ability, especially when both sections draw from the same cohort. The likely causes are inter-rater reliability failure and construct-irrelevant variance.

• Vague descriptors. 'Sections are clean' is not observable. Two graders can both be honest and disagree, because the criterion has no anchor.
• Severity and leniency effects. A is a harsh rater, B a lenient one; neither is necessarily wrong in intent.
• Halo effects. 'Doesn't listen' is a general impression that can leak into unrelated rubric lines.

Calibration process:

1. Rewrite the rubric as an analytic scale with observable anchors: part width within 1 cm, elevation consistent at 90 degrees, no visible ledge at the blend line, nape clean of stray hair.
2. Build an anchor set from anonymized student work or video, with one sample at each score point.
3. Have both instructors score five samples independently, then compare line by line and discuss every discrepancy larger than one point.
4. Set an agreement target, for example exact agreement on 80% of lines and within one point on all of them, and re-calibrate at the start of each term.
5. Where a line cannot be made observable, delete it or replace it.

Separate the three issues. Mixing them is the fastest way to lose the appeal and damage trust.

The score appeal: Re-score Jordan's practical against the calibrated rubric, using the retained work, photos, or video, by two instructors who did not originally grade it. If the score changes, correct the record and say so in writing. If it stands, explain the decision line by line with the rubric evidence. The appeal is about the work, not about whether Jordan is likeable or punctual.

Lateness: Handle it under the school's attendance policy as its own matter: notice, a meeting, documentation of the four missed demonstrations, and a plan to recover the missed content. Lateness does not retroactively make a haircut worse, and it cannot be used to justify a failing practical.

Clinic hours: Put a written make-up plan in place, supervised and documented, with real services and real clock time. It must not be papered over or traded for a passing practical score.

Timeline: Tell Jordan in writing what will happen and when, and give the director a recommendation that addresses procedures, not personalities.

Document: the original and revised rubric, score sheets from both sections, anonymized score distributions, the anchor samples used in calibration, the appeal form and re-score results, the attendance record, the make-up hour plan, and a communication log with dates.

To Jordan: a written outcome with the rubric lines, the re-score result, the reasoning, and the next step for the lateness and clinic hours. Keep other students out of the conversation entirely.

To the director: a short memo with the reliability data, the calibration steps taken, the re-score result, the separate attendance action, and the recommendation that both instructors use the anchored rubric going forward.

To the instructors: calibration findings framed as a system issue, not a personal failing. Show the anchor samples and the agreement data, and agree on a re-calibration date. Confidentiality matters here: no discussion of Jordan's case in the break room, and no comparing students across sections.

One more point worth flagging to the director: a 20-point rubric with a passing cut score of 14 leaves very little room for rater noise. Consider a critical-error rule, where a safety violation fails the practical regardless of the point total.