Skip to content

Live inter-rater agreement: catching data drift as it happens

Pathwize ResearchData quality monitoring2 min read
InsightPathwize AI
Research

Waiting for eval regressions to reveal bad data is too late. Live agreement between reviewers flags drift while you can still fix it.

Most quality problems in human data are not sudden, they drift. Guidelines get interpreted differently, a new cohort onboards, a task gets harder. By the time it shows up in model evals, the damage is already in the training set.

Why live agreement works

Measuring how often independent reviewers agree, in real time, turns quality into a signal you can watch rather than a report you read later. A drop in agreement is an early warning that a task or a cohort needs attention.

Agreement is a signal, not a verdict

One caveat: low agreement does not always mean low quality. Sometimes a task is genuinely ambiguous, or the guideline is unclear, and reasonable reviewers disagree. That is useful information too: it tells you to fix the instructions, not the people. High agreement on a task that should be hard can also be a warning that reviewers are colluding or guessing alike.

Act on the signal

With per-batch trust scores, you can pause a suspect batch, re-clarify the guideline, or re-route the task, all before it contaminates a dataset. Quality becomes something you steer, not something you audit.

Where it fits with gold-standard checks

Live agreement pairs well with occasional gold-standard items, questions with a known correct answer seeded into the workflow. Agreement catches drift and ambiguity in real time; gold questions calibrate against ground truth. Together they give you both a live pulse and an absolute reference.

Watch it on your pipeline

Pathwize surfaces live agreement and per-batch trust so drift is caught early. Book a demo to see it on a sample of your work.

Frequently asked questions

What is inter-rater agreement?+

It measures how often independent reviewers give the same judgment on the same items. Tracked live, a drop in agreement is an early warning that a task, guideline or cohort needs attention.

Why is live agreement better than spot checks?+

Spot checks look backwards and sample randomly. Live agreement turns quality into a signal you watch in real time, so you can pause a batch or fix a guideline before bad data reaches training.

Does low agreement always mean low quality?+

No. It can mean the task is genuinely ambiguous or the guideline is unclear, in which case you fix the instructions, not the people. Pairing agreement with gold-standard items helps separate the two.

Related reading

See Pathwize on your own data

Source verifiable expert data with provenance built in, EU-native and audit-ready.

Book a demo
← All posts

Related stories

Pathwize AIResearch

Trust scores belong on data batches, not just people

Why per-batch trust and live inter-rater agreement beat after-the-fact spot checks for catching quality drift.

ResearchPathwize AI

Expert evaluation vs crowd labelling for frontier models

Crowd labelling scales cheaply but breaks on hard, high-stakes tasks. Here is when expert evaluation pays for itself and how to combine the two.

ResearchPathwize AI

Human evaluation for reasoning models: what breaks

As models reason better, evaluation gets harder, not easier. Here is why generic raters fail on reasoning and what to do instead.