Consent screens that inform, not just cover.

Strip away institutional boilerplate and every consent framework asks for the same seven disclosures: who is running the study and where; what participation involves and roughly how long it takes; that it is voluntary and what stopping means in practice; foreseeable risks and any benefits, stated without spin; compensation and its conditions; privacy — what is collected, what is not, who sees it, how long it is kept; and a contact for questions plus one for complaints that is not you.

The writing standard is a cover letter, not a contract. Short paragraphs, labeled sections, reading level around eighth grade. Walls of legalese produce the thing review boards fear most: participants who consented without reading — which is to say, without being informed.

Replace every bracket, delete what doesn't apply, and run the result past your review board — their template beats this one wherever they conflict.

You are being invited to take part in a research study about [topic],
conducted by [name] at [institution]. Please read this page before deciding.

WHAT'S INVOLVED — You'll answer a survey taking about [N] minutes. There are
no right or wrong answers. [Mention anything unusual: audio, images, sensitive
topics.]

VOLUNTARY — Taking part is entirely voluntary. You may stop at any time by
closing the page; only submitted answers are kept. Because responses are
anonymous, we cannot identify and remove your data after you submit.

RISKS & BENEFITS — [State honestly. e.g. "No risks beyond everyday life are
anticipated. Some questions concern X, which some people find sensitive; you
may skip any optional question."]

COMPENSATION — [e.g. "You will receive £X via Prolific on approval."]

PRIVACY — Your responses are anonymous. We do not collect your name, email,
or IP address. [If using a panel: "Your panel ID is recorded solely to manage
payment and prevent duplicate participation."] Data will be [stored/shared]
as follows: [retention & sharing statement].

CONTACT — Questions or concerns: [researcher email]. For your rights as a
participant: [ethics board contact].

By selecting "I agree" below, you confirm you are at least [18] years old,
have read this information, and consent to take part.

Two lines in there earn their keep. The withdrawal sentence tells the truth about anonymity — you cannot un-submit an anonymous row, and pretending otherwise is a promise you can't keep. And the panel-ID sentence discloses the one identifier a panel study actually records, with its narrow purpose. Honest and specific reads better to participants and reviewers than vague reassurance.

Implementation shapes whether the consent is real. The screen comes before anything else — before demographics, before the welcome pitch warms them up. Agreement is an explicit action, not a pre-ticked box. And declining must do exactly what the screen promised: record nothing. A consent flow that quietly saved the decline event alongside a panel ID already broke its own first promise.

On panels, decliners also need the loop closed — route them to the screen-out exit so the submission resolves and the panel's compensation rules apply, rather than leaving them stuck.

In Plumeform, consent is a built-in gate: paste your text, and the survey shows agree/decline before question one. Declining stores nothing at all — no row, no partial, no event — and the study link, estimated duration, and screen-out routing all come from the same settings you already configured for the panel. The researcher tour shows it in place.

Sensitive topics deserve a specific mention in the risks paragraph and, where appropriate, skippable questions plus a support-resources line on the closing screen. Minors generally require parental consent plus child assent — a different protocol, not a reworded screen; panel studies typically require participants to be adults. Deception designs (including manipulated conditions participants aren't told about) need whatever your board requires — usually a debrief screen that explains the manipulation after submission. A/B message tests in market research don't rise to that bar, but if you are close enough to wonder, that is a board question.