The Two-Test Framework: How to Know — With Precision — Whether What’s Happening to You Is Real
%20(1).png)
Something is happening. You feel it every time. After every interaction with this person, something shifts — not in the room, not in anything you can point to, but in you. You feel smaller. You feel less certain of yourself. You feel like the ground beneath your own judgment has moved.
And then the institutional language arrives. You are told: you are too sensitive. No one else has complained. That is not what happened. He did not mean it that way. Everyone has bad days.
Each of these sentences does the same thing. It tells you that your internal state — the thing you are actually experiencing — is not evidence. It is a misperception. And the longer you are told this, the harder it becomes to trust what you know.
There is a framework that solves this. It was not designed for institutions. It was designed for the person inside the system — the person who has a feeling but not yet a method to confirm it.
The core idea, stated plainly: Two tests. Test One asks whether the target feels worse about themselves after an interaction. Test Two asks whether the venom is directed at people who are less powerful. Neither test requires you to prove intent. Neither test requires anyone else to confirm what you are experiencing. They do not replace the institution’s obligation to act. They give you something the institution cannot take away: a method to know, with precision, whether what is happening to you is real.**
The Condition the Tests Were Built For.
Robert Sutton did not design the two-test framework in the abstract. He designed it out of a specific experience — an experience in which he was hurt in a way that was structurally impossible to prove to anyone else. A colleague whispered something cruel inside a hug, at a moment when everyone around them saw only warmth. No one else heard it. No one else witnessed the shift from joy to damage. The evidence — by any institutional standard — did not exist. Only the effect existed. And the effect was precise: Sutton went from the happiest he had ever been about his work to doubting whether the achievement that had made him happy was even real.
The tests were built for this condition. Not for the loud, visible cases — not for the boss who screams in meetings or the CEO whose behavior makes the newspaper. Those cases are documentable. They can be witnessed. They can be reported and investigated and, in theory, acted upon. The tests were built for the cases that cannot be documented in any form the institution recognizes. The whisper. The tone. The comment that lands precisely and leaves no mark anyone else can see.
This is the condition that defines the audience for this framework: people who are experiencing something real, who know it is real because of what it does to them, and who have no institutional language to name it. The tests are not for the institution. They are for the person inside it — the person who needs a method, not a feeling, to confirm what is happening.
Test One. The Target’s Internal State.
Test One is stated simply: “After talking to the alleged asshole, does the ‘target’ feel oppressed, humiliated, de-energized, or belittled by the person? In particular, does the target feel worse about him or herself?”
Read the test carefully. Notice what it does not ask.
It does not ask whether the perpetrator intended to cause harm. It does not ask whether the person meant what they said. It does not ask whether anyone else witnessed the interaction. It does not ask whether the behavior meets any institutional threshold for “abuse” or “misconduct.” It asks one question: after the interaction, does the target feel worse about themselves?
Test One does not ask about the perpetrator. It asks about the target. That is the radical move in the framework — and it is the move that makes it work for someone inside the system. Every institutional dismissal of workplace harm begins with a question about the perpetrator: Did they mean it? Were they having a bad day? Is this part of a pattern? Test One begins somewhere else entirely. It begins with the person who was hurt. It says: your internal state is the first piece of evidence. Not the only piece. But the first.
The institutional sentence that most directly contradicts this test is “he is just joking” — the response that tells the target their internal state is a misreading of what actually happened. Test One says: it does not matter what the institution decides the interaction meant. What matters is what it did. If the target feels worse about themselves afterward — genuinely oppressed, humiliated, de-energized, belittled — the test is met. The institution’s interpretation of the perpetrator’s intent does not change the target’s internal state.
What “Feel Worse About Yourself” Actually Means.
The test is precise about the kind of damage it measures. It is not annoyance. It is not mild discomfort. It is not the ordinary friction that comes from working with people you do not like. It is a specific internal shift: the target feels worse about themselves than they did before the interaction.
Sutton applies the test to his own experience. After the colleague’s whisper at the graduation ceremony — an interaction that lasted less than a minute — he did not simply feel annoyed or embarrassed. He went from feeling the happiest he had ever been about his work performance to worrying that his teaching award would be taken as a sign that he was not serious enough about research. The award — which had been evidence of genuine effort and genuine improvement — became, in the space of a whispered sentence, evidence of his inadequacy.
That is what Test One measures. Not a bad moment. A shift in how the target regards themselves and their own work. The interaction did not change what Sutton had accomplished. It changed how he understood what he had accomplished. And that change — produced in under a minute, by a single sentence spoken inside a hug — is exactly the kind of damage that institutional language cannot accommodate. No one else saw it. No one else heard it. But it was real. And Test One confirms it as real without requiring anyone else to confirm it first.
Test Two. Power Direction.
Test Two asks: “Does the alleged asshole aim his or her venom at people who are less powerful rather than at those people who are more powerful?”
Like Test One, read it carefully for what it does not ask.
It does not ask whether the person is cruel in general. It does not ask about their character, their history, or whether they have ever been kind. It asks about direction. Where does the venom go? The same person can be warm, attentive, and collegial with people above them in the hierarchy — and systematically cruel to people below. Test Two does not care about the warmth. It asks: at whom is the behavior aimed? If it is aimed at people who have less power — people who cannot retaliate, who cannot refuse, who have less institutional protection — the test is met.
This is the test that makes the behavior structurally verifiable without requiring anyone above the target to witness it. The direction is observable by everyone at the same level. Multiple people can confirm it independently. And it removes the perpetrator’s most common defense — “I treat everyone the same” — because the test specifically asks whether the treatment differs based on power. Test Two is structurally a power-direction detector — the same detector that institutions refuse to apply during hiring and promotion, despite the fact that it is the only dimension where the pattern does its damage.
Why Both Tests. Why Not One.
Each test alone is insufficient. Together they form a diagnostic tool. This is not two questions asked independently. It is a designed system — and the logic for why each element exists becomes visible only when you read them together.
Test One alone would catch too much. A single bad day. A misunderstanding. A moment of rudeness from someone who is themselves under pressure. An interaction with a colleague who is going through something difficult and takes it out on the people around them. All of these could produce the feeling of being worse about yourself afterward. Test One does not, on its own, distinguish between the person who is having a bad day and the person who is systematically targeting you. It confirms that harm occurred. It does not confirm that the harm is structural.
Test Two alone is not sufficient either. Power-direction tells you that the behavior follows a pattern — that it is aimed consistently at people who cannot defend themselves. But it does not tell you whether the target is actually being damaged. Some people are blunt with everyone below them and produce no lasting harm. A manager who is curt and direct with subordinates but whose subordinates do not feel worse about themselves afterward has met Test Two but not Test One. The structural observation exists without a human cost attached to it.
Together: Test One confirms the harm is real. Test Two confirms the harm is structural — directed at people who cannot defend themselves. Both must be met for the framework to apply. This is what makes it a diagnostic tool rather than a grievance mechanism. It is not designed to validate every uncomfortable feeling. It is designed to identify the specific condition where a real internal shift has been produced by behavior that is aimed at people who have no institutional protection.
What the Tests Do Not Require.
This is the section that matters most for the person inside the system. Not what the tests ask — but what they do not require.
They do not require proof of intent. The perpetrator does not need to have meant to cause harm. Test One does not ask about meaning. It asks about effect. The effect is sufficient.
They do not require witnesses above the target in the hierarchy. The covert attack — the whisper, the tone, the comment that lands precisely — does not need to be seen by someone with the authority to act on it. Test One confirms the harm from the inside. The target’s own internal state is the first evidence.
They do not require a visible pattern — though a pattern strengthens the case significantly. A single interaction that meets both tests is sufficient to begin. The temporary versus certified distinction — whether the behavior is isolated or persistent — is a separate evaluation that follows after the framework has confirmed that something real is happening.
They do not require the perpetrator to admit anything. They do not require HR to investigate or act. They do not require the institution to change its position. The tests are not designed to compel institutional action. They are designed to give the target a method to confirm what they are already sensing — independently, precisely, and without requiring anyone else’s validation.
What This Looks Like in an Indian Office
Before the tests, there is a feeling. In Indian offices, that feeling has a specific shape — the shape of someone who suspects something is wrong but has no language to name it, and no confidence that naming it would be believed.
“Mujhe hi lagta hai kuch galat hai.” It seems to me that something is wrong. This is the sentence the junior employee says to themselves — not to anyone else. Not to HR. Not to a colleague. To themselves. Because saying it to anyone else would invite the response that the series has already named: you are too sensitive. No one else sees it. That is not what happened.
The tests do not replace this feeling. They answer it. They take “it seems to me that something is wrong” and give it a structure that does not depend on anyone else confirming it. Test One: after the interaction, do you feel worse about yourself? Not mildly uncomfortable. Genuinely oppressed, humiliated, de-energized, belittled. A shift in how you regard yourself and your own work. Test Two: is this person’s behavior aimed at people who have less power? Not at peers who could push back. At people who cannot.
If both answers are yes — the framework has confirmed what the feeling already knew. The junior employee does not need the institution to validate it. The tests have already done that.
From Feeling to Knowing.
The tests are not a complaint mechanism. They are not designed to make the institution act. An institution that does not want to act will not act regardless of how precisely the target can name what is happening. The tests do not change institutional incentives. They do not compel investigation or intervention.
What they do is something quieter — and, for the person inside the system, something more fundamental. They move the target from feeling to knowing. From “something seems wrong but I cannot prove it” to “I can name what is happening, I can name why it is happening, and I can name the structural conditions that make it possible.”
That movement — from feeling to knowing — does not require the institution’s permission. It does not require anyone else to agree. It requires only the two tests, applied honestly, to the interactions that are already occurring.
For someone inside a system that has spent years telling them their feeling is wrong — that they are too sensitive, that no one else sees it, that it is not what happened — knowing is the first thing that changes. Not the institution. Not the perpetrator. The target’s own certainty about what is real. That is what the framework provides. And that is why it was designed not for the institution, but for the person the institution has been trying to convince does not have a valid experience at all.
Source material: The No Asshole Rule — Robert I. Sutton (2007)