SJT Does Not Measure Whether You Are a Good Person

The reasoning most applicants use to skip SJT preparation goes: it measures character, and character cannot be crammed.

The first half is wrong, which makes the second half irrelevant.

SJT measures whether your judgement matches a standard produced by panels of clinicians and educators — a standard with specific and consistent priorities, most of which nobody has ever explained to you. Decent, thoughtful people underperform on it routinely, because being a good person and correctly predicting a professional consensus are different skills.

How it sits in your application

SJT is reported separately from the three cognitive sections, usually in bands rather than on the 300 to 900 scale. It is worth 69 marks.

Universities use it very differently: some as a threshold, some folded into a total, some barely at all. Check how each of your target schools treats it, because that determines how much of your time it deserves. If all your preferences treat it as minor, a few hours of calibration is enough. If one uses it as a hard threshold, it needs real attention.

Partial credit is generally available, which means two things. Never leave an item blank. And when torn between two adjacent options, either will likely earn something — so stop agonising and move.

The priorities the standard reflects

Learn this ordering and you have done most of the work.

1. 1**Patient safety first, always.** If an option prevents immediate harm it rates highly, however awkward it is socially.
2. 2**Stay inside your role.** Options where a student handles something alone that warranted a senior are rated poorly, even when well-intentioned. This is the biggest single source of error for candidates who over-value decisiveness.
3. 3**Gather information before acting.** Checking facts, asking the person, clarifying a misunderstanding — consistently rated above acting on assumption.
4. 4**Talk to the person first**, where it is safe and appropriate. Going over someone's head without speaking to them usually rates below a direct conversation. Unless there is immediate risk, in which case escalate.
5. 5**Honesty, without exception.** Concealing errors, misleading patients, covering for a colleague — rated poorly every time.
6. 6**Never do nothing.** Options involving ignoring it or hoping it resolves are almost always inappropriate.

Where your instincts will betray you

Four that feel reasonable and are marked down.

Loyalty to a friend. Protecting a colleague from consequences rates poorly when patients or standards are affected. The standard puts patient welfare above collegial loyalty, consistently.

Excessive independence. Handling something serious alone to avoid troubling a senior reads as a safety failure, not initiative.

Instant escalation. The mirror error. Formally reporting someone without speaking to them first, where the situation allowed it, reads as disproportionate.

Confrontation. Direct is good. Accusatory is not. Options phrased as challenges score below options phrased as questions.

Preparing efficiently

Not a big time investment. Not zero either.

Do a set and check your calibration — most people are systematically biased in one direction, either too harsh or too generous, and knowing which way you lean is half the fix.

Then read the explanations rather than the answers. The reasoning behind the official rating is the entire transferable content of this section. The answer alone teaches you nothing.

Keep asking what this specific person is permitted and expected to do. Most scenarios put you in a junior role, and the role boundary decides more items than anything else.

Two reading habits

Answer for the person named in the scenario. What is appropriate for a consultant is not appropriate for a first-year student.

And judge each option independently. In appropriateness questions the options are not ranked against each other — two can both be very appropriate. Candidates who assume the answers must spread across the scale force themselves into wrong ones.

We keep a free SJT mock available if you want to check your calibration before deciding how much time this deserves.