People who knew me in the laboratory tend to ask the same question when they find out what I do now, and they usually ask it with a slight wince. It comes out in different forms, but underneath it is always this: how do you take a mind trained to follow the data into a world that runs on assertion, repetition and tribal loyalty?
It is a fair question and I have been asked it enough times to have stopped giving the easy answer. The easy answer is that I bring rigour to a field that lacks it, which is both smug and not quite true. The honest answer is that the tension is real but it sits in a different place from where people expect, and that the Conservative tradition, properly understood, is far more hospitable to scientific habits than its critics or frankly some of its friends assume.
Start with what I mean by scientific thinking, because the phrase gets used to mean almost the opposite of what it should. It does not mean deference to experts. It does not mean deference to institutions with the word science in their name. It certainly does not mean treating consensus as truth, since the history of science is substantially a history of consensuses being overturned by somebody irritating. What it means is a specific discipline: you form a view, you say in advance what would make you abandon it, you go looking for that evidence rather than only for the comfortable kind, and you give up the view when the data tells you to. It means living with uncertainty rather than resolving it prematurely. And it means noticing the gap between what a study shows and what it is being used to claim, which in my experience is roughly the width of a press release.
I can tell you how it feels when that discipline bites, because it is not an intellectual experience. It is a bad afternoon. I have had a hypothesis I loved, one that performed beautifully in the data we developed it on, fall apart the moment it met an independent set from somewhere else. Nothing had gone wrong procedurally. The idea was simply not as good as I wanted it to be, and the only honest thing to do was say so to a room full of people who had spent money on the basis of my optimism. Science makes you do that reasonably often. It is a useful humiliation and I would recommend it to anybody entering public life.
None of these habits are uniquely scientific, and I would not claim otherwise. Any serious person in any field should aspire to them. They are simply better developed in scientific training because the feedback is harsher and arrives sooner. A trial eventually tells you whether the drug works. Politics almost never delivers a verdict that clean, which is exactly why bad thinking survives in it for decades, comfortably, drawing a salary.
So why do I file this under Conservative rather than technocratic? Because the technocratic version of evidence-based policy, which certainly exists and has plenty of adherents, holds that complicated social problems should be handed to credentialed people who work out the optimal answer and then implement it over the heads of the messy democratic process. That is not my argument. It is close to the thing Hayek spent a career warning about, and he was right to. The knowledge required to plan a society is not the kind of knowledge that can be gathered in one place, and pretending otherwise produces confident failure at scale.
My argument runs the other way. The Conservative tradition, from Burke through Oakeshott to the best of what the party thinks today, is founded on suspicion of grand designs. It distrusts blueprints precisely because they assume a level of knowledge nobody has. It prefers institutions that carry accumulated practical wisdom to models that look elegant on a whiteboard. It judges policy by consequence rather than intention. Those are empirical instincts, whatever else they are. When Burke asked for a disposition to preserve and an ability to improve, he was not requesting stasis. He was describing exactly the posture a good researcher takes towards an existing body of work: this was built by people who were not fools, so understand why before you start pulling at it, and change it on evidence about what is actually wrong rather than on the basis of what sounds better in the abstract.
Where the tension really lives, and I want to be straight about this, is in the cost of changing your mind out loud. In science, revising your position on new evidence is the job and earns you credit. In politics it is filed as a U-turn and used against you. I have sat on a planning committee and gone into a meeting with a settled view, listened to residents who knew things about a street that no officer's report had captured, and voted differently from the way I expected to. That is the correct outcome and it is also, in the currency of politics, expensive. I have not found a way to make it cheap, and I have noticed in myself the temptation to state things more firmly than I hold them because firmness plays better. I try to resist it. I would not claim I always win.
The other thing politics has taught me, which the laboratory did not, is that the data is sometimes the thing that is wrong. A resident tells you a junction is dangerous. The counts say the traffic is light and the collision record is clean. On the numbers there is no problem. Then you go and stand there at half past three in the afternoon and the problem is completely obvious, and you understand that the instrument was measuring the wrong thing at the wrong hour. Anyone with a scientific training should recognise that as a measurement failure rather than a conflict between evidence and anecdote. Quite a lot of what gets dismissed as populism is really local knowledge that the official dataset is too coarse to see.
What I hope to bring to Westminster is this disposition applied to the areas where I have actually done the work: health policy, life sciences and the industrial strategy around it, and the governance of clinical AI. In all three the debate is conducted with more certainty than the evidence supports, and in all three that happens in both directions. The case for NHS reform is regularly damaged by advocates who oversell what a particular intervention can deliver. The case against AI in clinical settings is very often made by people who have not read the validation literature. Arguments about drug approval are held almost entirely between parties who have never taken anything through an approval process.
I am not claiming this makes me right. Expertise is wrong all the time, and it must always answer to democratic consent rather than the reverse. What I am claiming is narrower and, I think, defensible: the argument gets better when somebody in the room has done the thing. That is not a technocratic position. It is a Conservative one, because it insists that decisions be grounded in how things actually work rather than how we would prefer them to. It requires people in politics who are genuinely willing to be refuted, which is rarer than it should be. I try to be one of them, and I would rather be judged on whether I manage it than on anything I have said here.
I write here in a personal and council capacity.