Features
Part of 5 things worth knowing about online status and identity
Anonymous versus verified: comparing four models of online identity
Four online status and identity regimes compared: anonymous, persistent pseudonym, real name and verified, with what each protects and who it exposes.
The argument about whether people should use their real names online has been running for as long as there have been names to use, and it does not resolve. It does not resolve because both sides are correct about different populations, and neither side's evidence answers the other's concern.
This page sets out the four regimes a service can choose, what each does to behavior and to accountability, and why the sensible design question is not which regime is right but which one a given space can afford.
What to take away
- There are four regimes, not two: anonymous, persistent pseudonym, real name, and verified. They behave very differently.
- The persistent pseudonym does most of the work people credit to real names, because reputation needs continuity and not legal identity.
- Real-name policies transfer risk onto whoever has most to lose from being findable, which is never the person proposing the policy.
- Verification answers a narrow question and is routinely read as answering a broad one.
The four regimes
| Regime | What persists | What accountability rests on | Who it protects | Who it exposes |
|---|---|---|---|---|
| Anonymous | Nothing between posts | Moderation of content only | People at risk from being known at all | Everyone, from anyone acting badly at no cost |
| Persistent pseudonym | A reputation attached to a name you chose | Loss of a reputation you built | People whose offline life and online life must stay apart | Someone who can be tracked across a long history |
| Real name | A link to an offline person | Offline consequences | People who want to trade on their standing | Anyone whose safety depends on not being found |
| Verified | A claim by the service that you are who you say | The service's process | Readers, against impersonation | People who cannot meet the process |
Nothing in that table says one row is best. Each row is a trade, and which trade is acceptable depends on what happens to the people in the worst position in that space.
What persistence actually buys
The strongest argument for real names is that accountability requires a person to lose something. That is right, and it does not require a legal name.
It requires continuity: a name carrying a history, one you would rather not abandon, and one that cannot be cheaply replaced. Whether continuity can be exported is the subject of the online status and identity platforms comparison.
That is what a persistent pseudonym is. Reputation systems built on stable handles work, for the ordinary reason described under reputation system design: past behavior is attached to a future you care about.
Communities with pseudonyms often behave better than communities with real names, and that is not paradoxical. Pseudonyms in a small place are known, and being known to the people you are talking to is the accountability that matters. Being known to the whole world is a different exposure altogether.
Anonymity removes continuity. That is why people choose it when the risk of continuity is high, and why it is hard to moderate. Both facts hold at once.
The behavioral half is documented as the online disinhibition effect. Read it as a description of a range of effects, not a verdict on one.
Why the real-name argument keeps repeating
It repeats because the two sides are answering different questions.
One side asks how to reduce abuse in a large public space. Attaching an offline person to a post plainly raises the cost of abuse, and the argument is sound as far as it goes.
The other side asks who pays when identity becomes mandatory. Several groups do.
People for whom being findable is dangerous. People whose employment or family situation makes their opinions costly. People using more than one name for ordinary reasons. People whose legal name differs from the one they use. For them the policy is not a cost increase; it is an exclusion.
Neither party is arguing in bad faith and no amount of data settles it, because it is a distributive question and not an empirical one. The general shape of the discussion, and why arguments about these services so often stall in this way, is set out under critiques of social media.
Verification, and what it does not mean
Verification says the service checked something. Usually that something is narrow: that an account corresponds to a named entity, or that a payment cleared, or that a document was shown. It is a defense against impersonation, and it is good at that job.
Readers extend it to mean importance, accuracy, endorsement, or permanence, none of which is in it. That is not the readers' fault: a mark beside a name is a strong visual signal. The underlying claim appears nowhere the reader will see.
That gap between what a signal certifies and what it is read as certifying is the same one that runs through every counter and badge, as online status and identity sets out.
Verification also changes what it means when it is withheld, which is the part designers usually miss. Once most prominent accounts carry a mark, the absence of one becomes a statement about an account, and the service has created a second signal it did not intend and does not control.
Choosing a regime for a space you run
Four questions, in order.
- What is the worst thing that happens to a participant if they are identified? If the answer is serious, the regime is decided already.
- Does the space need memory? If people must be able to build trust over time, you need persistence, which means pseudonyms at minimum.
- Can you moderate content on its own? Anonymity is workable only where you can, and it costs more moderator time than any other regime.
- What are you certifying, if anything? If you add a mark, write down what it means and put that sentence where readers will find it.
A space that gets this wrong rarely fails loudly, but quietly loses people who could not afford the regime. The ones who remain do not notice, so the record of these decisions stays one-sided.
The stages a group passes through as its composition changes are described in detail under internet communities. Why the departed leave no trace at all is covered under internet culture archives.
Common questions
Do real names reduce abuse?
They change who commits it and at what cost. They do not remove it, and they add a category of harm, which is what happens to people whose identification is itself the injury.
Is a pseudonym anonymous?
No, and the distinction is the point. A pseudonym accumulates a history, and a long history is often enough to identify someone even when no single post is.
Can a service offer more than one regime at once?
Yes, and many do, by space rather than by account. It works when the boundaries are clear to participants and fails badly when a post can move between spaces without the author expecting it.
Why not verify everyone?
Because verification excludes whoever cannot complete it, and that population is never the one the policy had in mind. The exclusion is the cost, and it is paid by people who are not in the room when the decision is made.
