Notice the words and timing around one disagreement
Read five original frequency prompts about pace, generalisations, repeated points, countercriticism and interruptions, with an exact example and a voluntary next step.
Five numeric frequency ratings produce a separate mean; a larger value describes more reported occurrences, not worse character or a validated danger level.
What counts as a reported occurrence
This group asks whether your voice became louder or faster, whether you generalised, repeated a point, answered criticism with criticism or began before the other person finished. An occurrence is not an explanation of why it happened. You might have been rushed, misunderstood, excited or distressed; the form cannot distinguish those possibilities. It also cannot observe whether either person felt heard. Read the particular prompts alongside their summary rather than treating the group name as an identity.
An exact mean, with no reverse coding
Suppose the five responses are 4, 3, 2, 4 and 2. They sum to 15 and the mean is 15/5 = 3.0. All five items are forward descriptions of frequency. This group is not reversed into a healthy-relationship score or subtracted from repair attempts. The same mean can arise when different behaviours occurred, so it cannot tell you whether interruptions or broad claims were most relevant. Inspect the individual statements privately if you want that distinction.
Replace one broad claim with an observable event
Imagine an ordinary disagreement about a changed plan. A possible sentence is: “When the plan changed this afternoon, I started repeating myself. I want to explain what I had understood before we decide what to do.” It identifies your action and the event. A counterexample is saying “You always make me shout” and presenting this test as proof. The form cannot assign responsibility for another person’s actions or establish that your account is the only accurate one.
Frequency depends on what had an opportunity to happen
A quiet month with one brief disagreement is different from a month containing many complicated conversations. The answer scale refers to relevant occasions, but the form does not count their number, duration, stakes or accessibility. A never response means the described action did not occur when you could judge it; no relevant opportunity belongs in the unscored option. Neither choice establishes better communication. Differences in memory, language and conversational expectations can also affect a report.
Four observation areas are not a sequence or ranking
The other groups describe expressing a pause, returning to a specific issue and checking a repair. These are editorial areas for reading, not verified independent factors or stages everyone passes through. Similar means do not prove the same underlying process, and a higher repair mean does not erase a harmful statement. The form makes no largest-group fight type. The separate conflict-styles questionnaire concerns approaches to competing concerns and solutions; its five-strategy scores cannot be converted into these four means.
Leave unavailable information unscored
Any unscored item leaves this group without a mean because all five numeric responses are required. Missing is not zero, an assumed middle value or evidence of avoidance. Other complete groups retain their own summaries. All twenty recognised unscored choices can finish the reflection and every full guide remains available. You do not need to reproduce an argument, ask a partner for a rating or recall a distressing event to make the numbers complete.
Research background does not validate these prompts
Research on conflict recovery has used observed couple interactions, including a discussion followed by a cooling-down task. These new self-report items do not reproduce that procedure or inherit its associations. FermatMind has not established reliability, norms, language equivalence, diagnostic cutoffs or breakup predictions for this edition. A four-week window helps organise recall and is an editorial choice. The form cannot determine who is right, identify abuse from a mean or decide whether a relationship should continue.
Choose one manageable observation
If you want to reflect further, privately choose one ordinary occasion and separate the event, your words and what you wished to communicate. You may decide to slow one response or ask whether another conversation is welcome. A discussion is optional and a refusal needs no test-based justification. If speaking would feel unsafe, stop the exercise. No result requires contact, an apology, forgiveness, reconciliation or remaining near a person who threatens you.