Why a signature at the top beats one at the bottom

An insurance form asks for the declaration of accuracy to be signed before the mileage figures rather than after. Reported mileage rises. The same form, the same signer, the same legal effect; only the position of the signature line changed, and with it the amount of driving people admitted to.

What everyone sees

Signatures are understood as verification, which happens after the work. That order feels logical: you attest to what you have written. It also means the attestation arrives when every figure has already been chosen, at which point it can only endorse decisions that are complete.

What is actually happening

Shu, Mazar, Gino, Ariely and Bazerman reported that signing an honesty declaration before rather than after reporting reduced misreporting in laboratory and field settings. The proposed mechanism is that a signature activates the signer’s moral self-concept, and the activation must precede the opportunity to shade a number in order to influence it. Mazar, Amir and Ariely’s self-concept maintenance account holds that people misreport up to the limit at which the behaviour would require revising their view of themselves; making that view salient early tightens the limit while the figures are still unwritten. It is worth noting that the original field result has been contested in replication work, and the laboratory effect is the better-supported part.

Why it stays hidden

The ordering effect hides because both versions feel equally honest to the person filling them in. Nobody experiences a bottom signature as a licence to shade, and nobody experiences a top signature as a constraint. The influence operates on which self-concept is active while the number is being chosen, and the signer has no access to that state — they see only a number that seemed about right.