Behavioral Metrics · 11 min read
Text Message Behavioral Metrics: What Textimony Quantifies
A practical framework for separating direct message counts from software candidates for threats, repeated contact, custody issues, boundaries, and other review questions.
The short answer
Message metrics fall into two groups that are easy to conflate and should not be. Direct measurements count what is plainly in the file: how many messages, sent in which direction, on which days, at what hours, with what gaps between them. These are arithmetic over the parsed record and they are as reliable as the export behind them. Software candidates are different — they are messages flagged against configured review categories, which is a suggestion for a person to examine rather than a finding. Textimony keeps the two separate deliberately, because a candidate count is not a count of proven incidents and presenting it as one would be a claim the software cannot support. Read direct measurements as facts about the file, read candidates as a queue for review, and treat any conclusion drawn from either as a human judgement that a person has to be able to explain. A metric is not a legal conclusion, clinical diagnosis, or safety plan. Category availability depends on the completed run, and each candidate should be checked against the message, surrounding context, participant direction, and available source fields.
A review framework—not a guaranteed output list
Abuse indicators — Threats, intimidation, coercion, monitoring, isolation, degradation, repeated fear-inducing contact. — Shows where message behavior may need legal, clinical, advocacy, or safety review without diagnosing the sender. Custody interference — Withheld parenting time, blocked child contact, exchange sabotage, school or medical information withholding. — Helps family-law reviewers build a chronology of observable parenting-access friction. Insults and degradation — Repeated demeaning labels, profanity used as a weapon, humiliation, contempt, appearance attacks, competence attacks. — Separates ordinary conflict from repeated identity-based or role-based verbal aggression. Threats and intimidation — Physical harm threats, reputation threats, legal threats, financial threats, immigration threats, self-harm pressure. — Surfaces severity, target, timing, and context windows for qualified review. Harassment and stalking cues — Repeated unwanted messages, contact after stop requests, location monitoring, showing-up language, third-party contact. — Connects repetition, boundary notices, fear language, and escalation over time. Boundary violations — Ignoring “do not contact,” showing up after being told not to, pushing access after refusal, using children or others as channels. — Shows whether stated boundaries were acknowledged, ignored, repeated, or escalated. Financial pressure — Rent, bills, support, car, job, immigration, housing, or property used as leverage. — Identifies coercive leverage and the surrounding messages that explain it. Evidence tampering — Requests to delete messages, change a story, hide screenshots, recant, lie, or stop documenting. — Highlights messages that may affect preservation, credibility, or litigation strategy. Escalation windows — Increasing frequency, severity, late-night bursts, post-separation surges, threat-after-boundary patterns. — Turns a long export into time-bounded review windows instead of isolated screenshots.
Who uses these metrics
Legal professionals can use source-linked metrics to triage large records, find issue windows, prepare authentication questions, and decide what needs deeper review. Therapists, clinicians, and advocates can use client-supplied records as structured context while keeping clinical judgment, mandated reporting, and safety planning separate from Textimony output. Consumers and survivors can use metrics to organize what happened, preserve context, and create review packets for counsel, advocacy, mediation, or personal records. Parents and co-parents can separate custody-interference records from general relationship conflict so parenting-time, exchange, school, medical, and communication issues stay visible. Review teams can compare counts, severities, date ranges, participant roles, and confidence notes instead of arguing from memory or selected screenshots.
Why source-linked metrics beat vibe labels
A label such as “abusive,” “harassing,” or “interfering” can hide more than it explains. Useful review language identifies the exact words, date range, direction, candidate category, and surrounding messages while keeping the legal or clinical conclusion separate. Clear category language helps a reader understand what was observed, inspect the relevant messages, and see where software output stops and human judgment begins.
How Textimony calculates carefully
Validate one supported two-participant export before analysis begins. Normalize messages while retaining the source identifiers, timestamps, sender labels, and text the export actually supplies. Keep direct counts, rule candidates, model candidates, and reviewer decisions separate. Use candidate records and selected excerpts to return to source context; daily chart points remain aggregates rather than citations. Keep uncertainty visible when sender identity, missing media, timezone, deleted content, or file conversion constraints affect interpretation. Separate observed message behavior from legal conclusions, diagnoses, admissibility opinions, and safety recommendations.
What behavioral metrics can Textimony quantify in text messages?
It reports message and daily-direction counts, and can surface rule or model candidates across several review categories where those are configured and the run completed. The caveat matters more than the list: a candidate count is a count of messages the software flagged for a person to read, not a count of proven incidents. Which categories are available depends on the completed run.
Is a behavioral metric the same as legal proof?
No. A metric counts something observable — how often, at what hour, in which direction — and counting is not proving. It cannot reach intent, credibility, or whether conduct meets a legal standard, and a high count may have an innocent explanation the messages do not contain. Metrics are useful because they make a long record navigable and give a reviewer somewhere to look, not because they establish anything by themselves.
Can therapists use Textimony behavioral metrics?
They can use client-supplied records as context, which is often more reliable than reconstructing a timeline from memory in session. What the metrics cannot do is diagnose, assess risk, or substitute for clinical judgment — they describe message patterns, not people. Therapy, diagnosis, crisis response, and any decision that follows from them stay with the professional and their own standards of practice.
Why not just use screenshots?
Because a screenshot captures a moment and these metrics describe a pattern over time. Repetition, timing, direction balance, and escalation are only visible across the whole record, and the screenshots people keep are by definition the ones that stood out — which makes them the least representative sample available. Metrics computed from a curated selection mostly measure the selection.
Published by
Textimony. Editorial status: Source-linked informational guide. Updated: 2026-07-12.
Sources
CDC NISVS FAQ: forms of intimate partner violence measured; CDC: Stalking and intimate partner violence; Technology-Facilitated Abuse in Intimate Relationships; Federal Rule of Evidence 901