RelationWatch Research Brief #09
A 374-case affair-during-marriage population, with conservative pre-discovery findings drawn from the evidence-rich cases whose chronology could be established.
Supported affair population: 374 casesHigh-confidence pre-discovery subset: 11Suspicion-only cases excluded from findingsThis Brief screened 2154 verified public cases and identified 374 cases with an affair supported as occurring during an ongoing marriage: 293 confirmed and 81 strongly reported/alleged. Another 152 cases were suspicion-only and 21 were unsupported or unclear; neither group contributes to the findings. Because 363 supported-affair cases did not document a sufficiently reliable before/discovery/after sequence, the ranked red flags use a nested 11-case high-confidence subset. Within that subset, Phone or message secrecy was observed in 8 cases (72.7%).
The 374-case population defines the broader context. The 11-case subset—not all 374 cases—is the denominator for ranked pre-discovery behaviors. This prevents discovery actions and post-discovery consequences from being mislabeled as warning signs.
What behavioral or situational warning signs appeared before discovery of extramarital affairs in ongoing marriages represented in RelationWatch?
| Total verified public cases searched | 2154 |
|---|---|
| Initial structured/full-text matches | 1667 |
| Not an ongoing marriage / later relationship | 1120 |
| Ongoing-marriage candidates | 547 |
| Unsupported/unclear affair evidence | 21 |
| Suspicion-only cases (excluded) | 152 |
| Supported affair, insufficient chronology | 363 |
| Qualifying cases | 11 |
| Distinct source URLs in qualifying audit | 11 |
| Incident-date coverage | 2020-01-01 – 2026-02-09 |
The chart describes all 547 ongoing-marriage candidates. Confirmed and strongly reported/alleged cases form the 374-case population; suspicion-only and unsupported records are excluded.
Categories can overlap; percentages use all qualifying cases as the denominator.
Composition of the fixed cohort, not a country comparison.
| Rank | Pre-discovery red flag | Cases | % of qualifying cases |
|---|---|---|---|
| 1 | Phone or message secrecy | 8 | 72.7% |
| 2 | Emotional or relationship distancing | 3 | 27.3% |
| 3 | Unexplained absences or schedule changes | 1 | 9.1% |
| 4 | Unusual closeness or hidden meetings | 1 | 9.1% |
Subpatterns are shown only when supported by the pre-discovery evidence retained in the audit.
| Messaging or social-media secrecy | 1 (9.1%) |
|---|
| Unexplained absences or schedule changes | 1 (9.1%) |
|---|
| Emotional or relationship distancing | 3 (27.3%) |
|---|---|
| Unusual closeness or hidden meetings | 1 (9.1%) |
| No financial category met the conservative chronology rule in this cohort. |
Role is reported only when a structured field or affair-context passage identifies it.
Discovery mechanisms are separate from pre-discovery warning signs and may overlap.
Examples are selected to cover the supported categories. The private screening audit retains the matched field, short evidence passage, chronology basis, classification, Case ID, country, and source references for every qualifying case.
Pre-discovery patterns: Emotional or relationship distancing, Unusual closeness or hidden meetings
Chronology support: Ordered event or chronology record
Pre-discovery patterns: Phone or message secrecy, Emotional or relationship distancing
Chronology support: Ordered event or chronology record
Pre-discovery patterns: Unexplained absences or schedule changes
Chronology support: Ordered event or chronology record
Pre-discovery patterns: Phone or message secrecy
Chronology support: Ordered event or chronology record
Pre-discovery patterns: Phone or message secrecy
Chronology support: Ordered event or chronology record
Pre-discovery patterns: Phone or message secrecy
Chronology support: Ordered event or chronology record
Pre-discovery patterns: Emotional or relationship distancing
Chronology support: Ordered event or chronology record
Pre-discovery patterns: Phone or message secrecy
Chronology support: Ordered event or chronology record
The population was generated from broad case-level full-text screening across narrative, article, transcript, source, event, and behavioral text, supplemented by existing structured fields; it did not rely on the affair drop-down value or the exact phrase “extramarital affair” alone. Search terms included affair, infidelity, cheating, adultery, lover, paramour, another man/woman, secret or illicit relationship, suspected affair, unfaithful, new partner, and third person. The search covered:
A case entered the 374-case supported population only when the record linked affair evidence to an ongoing marriage above suspicion-only. A case entered the narrower ranked-red-flag denominator only when a matching behavior was placed before discovery through explicit timing language or an ordered event/pathway record. The fixed snapshot and private evidence audit preserve the reviewable Case IDs and matching passages.