About Ties

How we talk to each other here

Ties puts people in one room who are used to very different rooms. A trial methodologist, a befriending co-ordinator, a ministry official and someone who has been lonely for three years are all reading the same thread. Almost everything that goes wrong in a network like this comes from forgetting that. These six norms are the ones we ask everyone to hold. They are light. Anyone can name one in a thread, nobody is in trouble for breaking one. If someone says “that was a well-actually”, the right response is “fair enough” and carry on.

1. Improve the discussion. Post when you are adding something: a finding, an experience, a correction, a question you actually want answered. “+1” is what the like button is for.
2. Argue with the claim, not the person. Disagreement is the point of a network like this one. That estimate does not support that conclusion is welcome and useful. You clearly do not understand methods is neither.
3. No feigned surprise. Nobody should be made to feel stupid for not knowing something: a practitioner who has not met preregistration, a researcher who has never seen a commissioning cycle. This place only works if people are willing to ask across those lines, and feigned surprise is the fastest way to stop them.
4. No well-actuallys. Correcting something tangential does not improve the post; it just raises the cost of posting. If the correction matters to the argument, make it. If it does not, let it go.
5. Practice knowledge counts as knowledge. “We ran this for four years and it did not work” is evidence of a kind, and often the only evidence there is. Say what kind it is, do not inflate it into a trial, and do not dismiss it because it is not one.
6. Write for someone reading in their second language. Short sentences, few idioms, acronyms expanded the first time. And nobody’s English gets corrected here unless they have asked for it.

Rules with consequences

The six above are norms. These are rules, and a moderator will act on them.

• Declare a commercial interest. If you sell something in this field, or stand to, say so on your profile. Nobody minds that you have an interest. People mind finding out later. This applies to consortium members and to the people running this platform, first of all.
• Effectiveness claims belong in the registry. If you want to say something works, put it in the Intervention Registry with its evaluation status, where the claim sits next to what backs it. A forum post is not the place to make a claim nobody can check. This applies to a grant-funded programme exactly as much as to a product.
• Post your own work the way a paper gets posted. A researcher announcing a study and a founder announcing a product are doing the same thing: telling people who might care about work they have a stake in. Ties holds both to the same standard, and the standard is about the post rather than about who wrote it. Four things make it a post rather than an advertisement.
  • Say what it is built on. What the thing does, what sits behind it, and where that evidence runs out. If there is nothing in the post a reader could disagree with, it is an advertisement.
  • Declare the interest in the post itself, not only on your profile. One line will do: this is ours, and this is how it is paid for.
  • Put the claim in the registry and link it. Anything of the form “it works” goes in the Intervention Registry next to its evaluation status, and the post links the entry. A grant-funded programme is held to this exactly as tightly as a product is.
  • Take the replies as review. “Your comparison group is wrong” and “that is a retention number, not an outcome” are the reason to post here rather than on LinkedIn. If you cannot answer that in the thread, it was a pitch after all.
  What is still removed: the same thing posted more than once, sales approaches by private message to members who did not ask for one, and anything shaped as a funnel, such as a demo booking, a waitlist or a discount code. Pricing and commercial terms are a conversation to have somewhere else. Moderators apply all of this most strictly to consortium members and to the people running this platform, who have the most to gain from it being applied loosely.
• Other people’s personal data stays out. Do not post identifying details of service users, participants or colleagues. Not names, not photographs, not anything that would identify someone in a small community. Anonymise the case before you tell it.
• What a member shares here stays here. Do not repost someone’s message, and especially not a personal disclosure, anywhere else without asking them first.
• This forum is not a recruitment pool or a dataset. Do not recruit participants, circulate surveys, or collect, quote or analyse posts for research without asking first. Post in General, say what the study is, who is running it, what your ethical approval and legal basis are, and let people opt in. Members should be able to write here without becoming data.
• Do not post embargoed or confidential material: unpublished results you do not have permission to share, draft policy, or anything under a non-disclosure agreement.
• Link papers, do not paste them. A DOI or a link is fine; a paywalled full text pasted into a thread is not.
• No harassment, discrimination or hate speech, including the subtle kind. Persistent belittling of a group, or of a person, is not a style difference.

A note on distress

This is a network about loneliness, and people will sometimes write about their own. That is welcome, and it is part of why lived experience belongs in this room. But Ties is not a support service and nobody here is on duty. Be kind, do not treat what someone has told you as material, and if you are worried about someone, say so gently and point them to help where they live. If you are in crisis yourself, please contact your local emergency or crisis service rather than waiting for a reply here.

What happens when something goes wrong

Almost everything is settled at the first step. The ladder exists so that the rare serious case is handled predictably rather than improvised.

1. A member names it. “That reads as a well-actually”, or “could you say that without the swipe?” Most things end here, and this is the version we would like you to reach for.
2. A moderator edits or hides one post and says publicly why. Editing is preferred to deleting: the thread keeps its shape and everyone can see what the standard is.
3. A private word. For a pattern rather than a post.
4. A temporary suspension, usually a week or two, for a repeated or serious breach after a warning.
5. Removal from the platform. Rare, and never a moderator’s decision alone.

Grounds for immediate removal, without the ladder: harassment of or threats towards a member; hate speech; deliberately publishing someone’s personal data; impersonating another person or organisation; sustained commercial spam after one warning; and covertly harvesting the forum for research or commercial data. These are the cases where the damage is done by the time a warning would land.

Appeals go to the Ties moderation team, to someone who did not take the original decision. Every moderation action above step two is logged with its reason, and any member can ask what the log says about them.

Moderation is done by people who have other jobs. If you see something, flag it rather than replying to it.

Where this comes from

The norms above are adapted from two sources worth reading in full: the Recurse Center’s social rules, which are the origin of no-well-actuallys and no-feigned-surprise and of the idea that a light correction is not a punishment; and Discourse’s rules of civilized discourse, which are where “improve the discussion”, “criticise ideas, not people” and “flag it rather than reply to it” come from. What we have added is the part specific to this network: the mix of research and practice, the second-language rule, the registry, and the line about not treating the people here as data.