Reflect on passing on unverified accounts, checking what may be shared and updating listeners after an error. A score proves neither motive, truth nor actual harm.
Reflect on passing on unverified accounts, checking what may be shared and updating listeners after an error. A score proves neither motive, truth nor actual harm.
Four valid ratings yield this part’s accuracy mean, without overall toxicity, bands or a personal verdict.
What this part explores
This part follows four points at which an account leaves your hands: whether checking precedes retelling, whether guesses are identified, whether shareable details are confirmed, and whether a correction reaches the original listeners. It does not count how much you talk or classify every discussion of an absent person as harmful. Relaying a confirmed message at someone’s request differs from passing a guess along as a fact. No names, original messages or event records are collected. Your mean therefore cannot establish what happened, why you spoke, or whether a particular account was true.
Reading accuracy ratings
A higher mean means that these four sentences fit your present recollection more closely. It does not establish more frequent retelling, greater harm, or an identity as someone who spreads rumours. Four ratings of three might reflect different practices across settings or uncertainty about the examples. That category enters the arithmetic without becoming a stable middle habit. A lower mean cannot certify that every message is reliable either: opportunities may have been scarce or easier cases more memorable. Notice which sentence was difficult to rate before treating a small decimal difference as meaningful.
An everyday example you can check
Suppose a companion says that an activity may have moved, and you are about to tell other participants. Before sending, separate an organiser’s confirmed change from the companion’s guess. Consider which details participants need and which private reasons need not travel with the message. If an old address has already been forwarded, checking it later changes your understanding but does not automatically update the original recipients. This example provides a chain of actions to review. It asks for no event details or chat records, and one missed correction does not determine a fixed result.
An optional phrase to try
If useful, try: “I have not checked this part, so I will not forward it as a fact. Which details may I share?” When correcting an earlier account, identify the message being corrected without adding unnecessary private material. The purpose is to make the information’s status clearer, not to prove that you never make mistakes. Immediate verification will not always be possible; leaving a claim uncertain or choosing not to circulate it can remain an option. You do not need to gather or upload someone else’s private material to complete this reflection.
How it differs from the other four parts
This part centres on checking an account and tracking where it travels. Comparisons concern how someone’s update is reframed; hints concern whether an objection becomes a clear request; explanation-first concerns the order of responding to a challenge; old-dispute carryover concerns the scope of the current discussion. More than one practice can occur in a conversation without becoming the same mechanism. The questionnaire has not established that its parts are interchangeable or additive units of harm. Read each separately, and do not use one part to fill information missing from another.
What this result cannot classify
Describing your own experience to an appropriate person when seeking help is not equivalent to circulating an unchecked rumour. Protecting private details is not automatically concealment intended to control someone. The purpose and recipient of information can differ substantially between ordinary conversation and situations involving support or safety, and these four items do not assess those conditions. A self-rating cannot decide who may speak, who should stay silent, or who bears responsibility for an event. It also cannot require contact with someone you do not wish, or are not in a position, to contact.
A small, bounded review
For a closer look, choose one ordinary event that can be reviewed without exposing anyone’s privacy. Separate what you knew at the time, what you guessed, what was confirmed later, and who received the account. No names or personal label are needed. Choosing to mark uncertainty, narrow the shared details or send a correction is a concrete action; it is not improvement established by this mean. Opportunities, recall and the events available for comparison may change between attempts. This edition provides no reliable-change threshold for deciding that a score difference reflects lasting change.
Recomputing five separate parts
All twenty items offer 1 Very inaccurate, 2 Somewhat inaccurate, 3 Uncertain or depends on the situation, 4 Somewhat accurate, 5 Very accurate, or a separate NA. Five parts contain four directly keyed items each, with zero reversals. A part needs all four numeric ratings: its raw sum of 4–20 divided by four gives a two-decimal descriptive mean and a rated count, never reconstructed from a one-decimal mean. Three is numerically valid without becoming stable neutrality, frequency or intensity. NA is not zero, three or an estimate from the remaining items. Missing information in one part does not erase another. All twenty selections finish the attempt, and all five full guides remain readable. There is no highest type, overall sum or mean, percentage, percentile or saintly-to-severe band. Averaging ordered categories is a display convention, not established interval precision, magnitude of harm or significant change.
Edition, full free guides and privacy
This new form contains sixteen newly authored items plus four Chinese items from one part of an owned mini-program preview with new English translations. It is neither a whole-form migration nor a validated scale. “Toxic trait” is a common search title; results discuss specific practices without judging anyone’s morality, diagnosis, danger, responsibility or career. Full text is free without another assessment. Drafts require opt-in, stay on this device for twenty-four hours and clear on completion, exit or retake; declining storage does not prevent completion. Sharing includes only the test name and public entry, without answers, personal means or missing counts. No private event material is requested, and no population rankings, language equivalence or local norms are established.