Critical Thinking, and How an Administration Builds It

An officer is rarely asked to be brilliant. He is asked to notice, before he signs, that the proposal in front of him rests on something nobody has checked. This part is about that habit — and about why an organisation has to work so hard to keep it alive.

Critical thinking carries an unfortunate reputation in government. It is heard as a synonym for scepticism, or for the officer with an objection ready before the file is untied. It is neither. Scepticism is a mood and contrarianism a posture, and both can be held without doing any work. Critical thinking is work: the discipline of separating what a proposal has demonstrated from what it has merely asserted, and holding the two apart long enough to see which is carrying the argument.

The syllabus places it among aptitudes rather than values, which is worth taking seriously. A value is held; an aptitude is exercised, which means it can be practised, coached, allowed to decay, or trained out of a person by an office that has no use for it. That is what the examiner probes when a question asks whether moral aptitude can be developed through training or is inherently possessed. The raw capacity is common; its exercise is not, and the difference is made largely by the organisation an officer works inside.

The three questions that do the actual work

The first is the plainest. What has this file shown, as distinct from what it has stated in a confident voice? A note opening with “it is well known that” asks to be believed rather than offering evidence. By the third desk an estimate has usually hardened into a figure and the figure into a fact. Reading critically means asking where each number entered the record and who is in a position to be wrong about it.

The second asks whether the stated purpose of a measure matches its likely effect. A scheme meant for the poorest households can be routed through a channel those households cannot enter, and the file will not say so, because a file records intention and not consequence. The third is the one most often skipped: who bears the cost of a decision that looks costless on the file? A deadline advanced by a fortnight costs the sanctioning authority nothing and the field staff a fortnight of nights. The three questions exist to surface the assumptions a proposal has buried.

Why the hierarchy punishes the questioner

Here is the difficulty. A hierarchy does not merely fail to encourage the habit; it suppresses it. The junior officer who questions a proposal is read not as careful but as obstructive, as slow, as someone who does not appreciate the pressure the section is under. The safest note he can write is the one that agrees. A dissent that proves wrong is remembered for years; an agreement that proves wrong is absorbed by the system with no name attached to it.

The asymmetry has predictable results. Groupthink is rarely the product of everyone agreeing. It is the product of everyone separately concluding that the cost of saying otherwise is not worth paying, and then reading the resulting silence as consensus. Cognitive bias compounds it: an officer who has already given a year to a project reads new information as confirmation, because the alternative is to have wasted the year. The file then acquires a momentum unrelated to its merits, and the meeting called to review it becomes a meeting to endorse it.

What follows should be read as an answer to that problem, not as a list of desirable activities. A culture of inquiry matters precisely because the default culture punishes inquiry. An officer will not improvise a challenge to his superior at the moment one is needed; he will make it only if the organisation has already given him a form in which such a challenge is ordinary.

Making inquiry ordinary rather than brave

An organisation that wants questions asked must lower their price. Discussion forums on live policy issues move argument off the file — where positions harden and are recorded — into a room where a view can be tried out and withdrawn. Monthly review panels make scrutiny scheduled rather than provoked: nobody has to decide to raise a problem, because the calendar has already decided. Feedback mechanisms give criticism a route that does not depend on the nerve of the person offering it.

Training and development programmes carry the same logic. Workshops on ethics and decision-making, policy simulations built on real constraints, and analysis of past decisions whose outcomes are known share one feature: they let an officer be wrong where being wrong is free. The better version assigns someone the devil’s advocate role explicitly, so that dissent becomes a duty rather than a personal risk. Problem-solving sessions extend the same permission to unfinished work — brainstorming on a current difficulty, a departmental think tank on a recurring one, an innovation lab for approaches that may not survive the district — so that a proposal is examined before anyone has invested standing in it.

Learning that does not stop at the probation year

Continuous learning is the least glamorous of these practices and the most easily reduced to a certificate. Its real function is to keep an officer in the position of a student long enough to remember what not knowing feels like, because a person who cannot occupy it will not be able to hear an objection either. Support for further study, courses in ethics, policy and analysis, and arrangements with academic institutions all serve that end. What matters is less the syllabus than the interruption: someone recently examined on a subject is harder to persuade that his view of it is complete.

Cross-functional learning attacks a different failure. A department reviewing its own work does so with its assumptions intact; the questions it cannot think to ask are the ones its neighbours would ask first. Inter-departmental task forces put the revenue view, the police view and the health view in one room, where each must explain itself to the others. Mentorship does this across time, peer review laterally — provided it is constructive rather than a ranking exercise. Exposure outside the service — exchange programmes, expert panels from academia and industry, community engagement — introduces a listener who owes the department nothing. That last is respected least and matters most: a person living inside a consequence names the unstated assumption in one sentence.

Tools, reflection and the frame that keeps them honest

Analytical technique gives all of this something to bite on. A SWOT assessment forces a policy’s weaknesses onto the same page as its strengths. Root cause analysis interrupts the reflex of treating the most visible symptom as the problem — the reflex that orders a transfer when a procedure needed changing. Impact assessment sets intended effect against observed effect, which is the second of the three questions put on a formal footing. None of these tools thinks for an officer; they only make it harder to skip a step.

Reflective practice turns experience into something usable, and its absence is why long service and good judgement are not the same thing. A decision journal records what was expected at the time of a decision and can be read afterwards against what happened; without it, memory quietly rewrites the expectation to match the outcome. An after-action review does this collectively after a major exercise, a debriefing for smaller ones, provided the exercise looks for the cause and not the person, since a review understood as an inquiry into blame produces accounts that are careful rather than true.

An ethical decision-making framework holds the rest in place. Scenario-based training, committees that examine proposals for their ethical implications and not only their financial ones, and a stated sequence for working through a dilemma make the ethical question a routine step rather than an act of individual courage at the end. Rigour without that frame produces an officer who is very good at justifying a decision he should not have taken.

What the examiner is actually testing

Questions here arrive in three shapes. The theoretical ones ask you to define an aptitude, explain how it contributes to effectiveness, or set out how two aptitudes relate. The case-based ones give you a scene and ask which aptitude was missing, or how a named one might be built. The integration questions ask how several aptitudes work together on a complex problem, which combination a particular service demands, or how the ethical and the practical pull against each other. The first rewards precision; the other two reward the habit described above.

The case section shows this most clearly. A weak answer announces a verdict in the opening line and spends the remainder defending it. A strong one does what an officer does with a file: it names the assumption the case has left unstated, sets out the alternatives with their costs and the people who will bear them, and only then commits. The marks are not for the decision. They are for the visible working — the steps by which a defensible conclusion was reached, which on a written paper is the only thing separating judgement from opinion.