When an Assessment Goes Wrong

Things fail. The response determines whether it becomes a problem with a process or a problem with your reputation.

Once the criteria described in “When an Assessment Goes Wrong” are fixed, the practical challenge is running them consistently across assessors. A coordinator can use the official explanation to plan review time, compare the effort spent at each stage and spot avoidable delays, without confusing administrative efficiency with the quality of the evidence collected from candidates.

For an independent benchmark, compare the process with WHO guidance on mental health at work; the useful test is whether the local method remains job-related, proportionate and explainable.

The common failures

A technical failure during a timed assessment.

An exercise or question set circulating publicly.

An assessor behaving inappropriately.

A scoring error discovered after decisions were communicated.

And a stage discovered to be producing a group difference nobody had measured.

The first day

Stop the affected process. Do not continue making decisions on material you now doubt.

Identify who is affected, specifically.

Tell them promptly, even before you know the remedy. A message saying you are aware and will confirm within two days is better than silence.

And preserve everything, including the configuration and timings, because the question of what happened and when will arise.

Technical failure during an assessment

Offer a re-sit without prejudice, promptly, with an explanation.

Do not score the affected attempt.

And do not require candidates to demonstrate the failure, which places an evidential burden on the people your system failed.

A leaked exercise

Assume everybody has seen it, because you cannot tell who has.

Replace it rather than trying to identify who used it.

And note that a behavioural question set circulating matters less than an exercise with a scoring key, so the response should be proportionate.

An assessor behaving badly

Remove them from the process immediately, pending whatever review applies.

Contact the affected candidates and offer a re-assessment with different people.

And look at their earlier decisions, which is the uncomfortable part and is necessary.

A scoring error after decisions went out

Correct it and tell everybody affected, including anybody who was rejected in error.

Whether to reopen an offer already accepted is a judgement, and the default should lean toward the wronged candidate.

Hiding it is the option that becomes a reputational event.

Afterwards

Write down what happened, what caused it, and what changed.

Check whether the failure was detectable earlier: most were.

And tell the affected candidates what changed, which costs a sentence and is disproportionately well received.

What distinguishes a good response

Speed, specificity, and not requiring candidates to prove your failure.

Organisations that handle this well are not the ones with fewer failures. They are the ones that stopped quickly and fixed it without argument.

What to check

Is there a named person who can pause a process?

Would you detect a technical failure, or rely on candidates reporting it?

Do you have a route for a candidate to complain about an assessor?

And has anybody rehearsed what you would say?

The point

Stop the process, tell affected candidates promptly, and do not require them to demonstrate your failure..

Underlying all of this

Almost everything in this collection reduces to one discipline: write down what the job requires, assess that thing directly, record the evidence, and look at your own outcomes afterwards. None of it requires buying anything, and organisations that do those four things consistently outperform ones running longer processes built from instruments chosen before the requirements were known.

The recurring pattern

The recurring failure across every section here is the same: measuring what is convenient rather than what matters, then never checking whether it predicted anything. The check is an afternoon of work once a year, and it is the step that separates a process that improves from one that merely persists.