The Unaligned Steering Wheel: Why Religious Confirmation Bias Is Larger

If you have driven a car whose wheels are out of alignment you know the feeling. The car pulls to one side. Not violently — you can drive it for years — but constantly, and if you take your hands off the wheel for a moment you drift toward the ditch. The only way to go straight is to apply a small, steady pressure the other way, and to keep applying it for the whole journey.

That is what arguing with a believer is like, and I want to explain why the pull is not symmetric.

Everyone Has the Bias

Let me concede the obvious first. Confirmation bias is universal. Everyone rates their own arguments more highly than they deserve, remembers the evidence that helped them and forgets the evidence that did not, and presents their case as stronger than it is. I do it. You do it. The theist across the table does it.

So the believer's reply is ready: you have a bias too, so we are even, and your confidence that I am biased is itself just your bias talking.

That reply would be correct if bias were a fixed quantity, the same for everyone regardless of what they believe. It is not.

The Bias Scales With the Prize

The strength of the pull is proportional to what the belief promises. That is the whole argument, and I think it is close to a law.

Confirmation bias is not a random malfunction. It is motivated reasoning, and the motivation is the thing the belief would deliver if true. The more a belief offers, the harder the mind works to protect it, the more it discounts contrary evidence, and the more it inflates the arguments in its favour. This is not controversial as psychology; it is the reason people are worse at judging their own investments than other people's.

Now compare the two prizes.

Religion offers eternal life. It offers reunion with the dead. It offers justice — every wrong eventually righted, every unpunished cruelty eventually punished. It offers meaning, a place in a plan, a person at the centre of the universe who knows your name. It offers that death is not the end. There is no larger set of promises available to a human being; the religions were assembled over thousands of years precisely by selecting for whatever people most wanted to hear.

Atheism offers a cold universe. Nobody is listening. The dead are gone. Justice is whatever we manage to build ourselves, and mostly we do not manage. Your life means what you make it mean and then it stops.

One of these positions is enormously appealing and the other is not, and the appeal is what powers the bias. The believer has every conceivable motive to overrate the arguments for God. I have almost none to overrate the arguments against — my position is not one anyone adopts because it feels good.

What Follows for the Argument

This does not tell you which side is right. A belief can be appealing and true. But it tells you where to put your thumb on the scale when assessing the two sides' self-assessments.

When a believer tells me an argument is decisive, I know that the same argument, pointing the other way, would strike him as feeble, and that his sense of its strength is being produced partly by what it would give him if it held. When I catch myself finding an argument against God compelling, I should check the same thing — but the check comes back with less to correct, because the reward for being right is so much smaller.

So the pressure I apply to the wheel is not a bias of my own. It is the compensation for a known and measurable pull, and it should be applied constantly, not once. Every argument for God gets a discount. Not because it comes from a believer, but because it is an argument for the most attractive conclusion ever formulated, and attractiveness is the one thing we know inflates confidence.

The Objection From Comfort

The usual reply here is that atheism has its own rewards: freedom from guilt, no religious obligations, the pleasure of feeling cleverer than the pious, membership in an intellectual tribe.

These are real, and they are small. Compare them honestly to you will live forever and see your mother again. The asymmetry survives the comparison by orders of magnitude. If you doubt that, notice which way the deathbed conversions run.

Driving Straight

I have argued elsewhere that consolation is not evidence — that how a belief feels tells you nothing about whether it is true. This is the practical sequel. Because the consolation is so unequally distributed, the bias it produces is unequally distributed too, and honest reasoning has to correct for the difference rather than pretend both sides are pulling equally.

The car pulls right. You steer a little left, all the way home, and you do not apologise for it. That is not prejudice against the road. It is knowing your vehicle.