OPERATIONS · RESILIENCE
What Happens When Your Only Assessor Is Off Sick? Building Marking Resilience in a Small Training Provider
When marking capacity is concentrated in one or two people, a single absence hits turnaround directly. What resilience looks like at small scale, without a second full-time hire.
By Eduface · July 2026 · 8 min read
Ask a training provider with two or three assessors what their marking plan is if one of them is unreachable for two weeks, and you’ll usually get a slightly uncomfortable pause before the answer, because the honest answer is often everything slows down and we hope it’s not for long. That’s not a criticism. It’s just what happens when marking capacity is concentrated in one or two people, which is the normal state of affairs for a small provider, not a sign anything’s being run badly.
Why this is a structural risk, not a bad luck problem
In a large institution, marking capacity is distributed across enough people that one absence barely registers. In a small provider, it’s the opposite: a handful of assessors, sometimes genuinely just one for a particular subject or qualification, are carrying the entire marking load for that area. That’s not inherently a weakness, it’s just what efficient staffing looks like at small scale. But it does mean a single absence, illness, a family emergency, someone finally taking the leave they’re owed, hits your turnaround time directly and immediately, in a way it simply wouldn’t at a bigger institution with more redundancy built in.
What the research on marking reliability actually says about workload
It’s tempting to assume that a tired, overloaded assessor is a straightforwardly less reliable one, and there is real research pointing that direction, but the picture is genuinely more nuanced than a simple story about fatigue causing sloppy marking. A well cited multilevel study of England’s national curriculum English writing test found no significant evidence of rater drift, meaning assessors marking large volumes over an extended period didn’t become systematically more lenient or severe as fatigue set in (Leckie & Baird, 2011). Other studies, and a wider literature review of marking reliability commissioned in England, point the other way, flagging fatigue and rater drift as real, documented risks under certain conditions (Meadows & Billington, 2005).
The honest summary is that the research doesn’t give you a single clean number to plan around. What it does tell you clearly is that marking reliability under sustained load is a genuinely studied problem, not an assumption, and that the risk is highly dependent on specifics: how well trained the assessor is, how the marking scheme is designed, how much volume they’re actually carrying. That’s worth knowing, because it means the solution isn’t simply reduce workload in the abstract. It’s building a structure that doesn’t concentrate all the risk in one person’s calendar in the first place.
What resilience actually looks like at small scale
You don’t need to hire a second full-time assessor to fix this. A few things matter more than headcount.
A documented marking scheme that isn’t only in one person’s head. If your primary assessor’s judgement calls exist mostly as tacit knowledge, I just know what a pass looks like, you have a single point of failure even when that assessor is perfectly healthy and reliable. Writing that judgement down, in enough detail that a second person could apply it consistently, is the single highest-leverage thing most small providers can do here.
A genuine second reviewer, even a part-time or occasional one. This doesn’t need to be a full second assessor’s worth of capacity. It needs to be someone who can pick up a batch of marking on short notice without starting from zero on your marking scheme and standards.
A clear, honest answer to learners about what happens during a gap. Providers that communicate proactively about a delay, even a short one, tend to weather it far better than providers that go quiet and let learners assume the worst. This connects directly to the point in our companion piece on why slow marking costs providers goodwill: it’s rarely the delay itself that does the damage, it’s the uncertainty around it.
AI-assisted marking as a genuine buffer, not a replacement. This is exactly the scenario where AI-assisted marking, properly built with mandatory human review, earns its keep. It doesn’t remove the need for a human decision on every submission, our companion piece on human-in-the-loop marking covers why that step has to stay real. But it does mean a stand-in reviewer, someone less immersed in your specific marking scheme than your primary assessor, has a structured, reasoned draft to work from rather than a blank page and a stack of unfamiliar submissions.
The plan is worth having before you need it
None of this is complicated, but almost none of it gets written down until after the first real disruption, when a provider that’s been running fine on one key assessor suddenly has to explain a two week delay to forty anxious learners. Writing the marking scheme down, identifying who your second reviewer would actually be, and having an AI-assisted workflow already tested rather than something you’re evaluating for the first time under pressure, are all things worth doing while everything is calm, precisely because they’re much harder to do well once it isn’t.
Frequently asked questions
Does research show that tired assessors mark less reliably?
The evidence is mixed rather than settled. Some studies find little evidence of fatigue-driven drift in large operational settings, others flag it as a real risk under certain conditions. The safer conclusion is that reliability under load depends heavily on training, scheme design, and volume, not that fatigue is either a non-issue or an inevitable problem.
Do I need a second full-time assessor to fix this?
Not necessarily. A documented marking scheme and a part-time or occasional second reviewer who can step in on short notice covers most of the risk without the cost of a second full-time hire.
Can AI marking tools actually replace the need for a backup human assessor?
No, and they shouldn’t try to. What they can do is make it much easier for a less familiar reviewer to step in during a gap, since they’re working from a structured draft rather than starting cold.
What’s the single most useful thing a two- or three-person team can do about this now?
Write down the marking judgement that currently lives mostly in one person’s head. That’s the step that turns we hope nothing happens into an actual plan.
Sources
Leckie, G. B., & Baird, J. A. (2011). Rater Effects on Essay Scoring: A Multilevel Analysis of Severity Drift, Central Tendency, and Rater Experience. Journal of Educational Measurement, 48(4), 399-418.
Meadows, M., & Billington, L. (2005). A Review of the Literature on Marking Reliability. Report for the National Assessment Agency, AQA Centre for Education Research and Policy.
Parasuraman, R., & Manzey, D. H. (2010). Complacency and Bias in Human Use of Automation: An Attentional Integration. Human Factors, 52(3), 381-410.
Give a stand-in reviewer a running start
Eduface hands a backup reviewer a structured, reasoned draft instead of a blank page. Book a demo or start free.