Emotional regulation signals — response latency patterns, communication consistency, escalation or de-escalation trajectories — are among the most predictive behavioral dimensions. They are also among the most dangerous to score without careful governance.
The Engines
- Emotional Regulation Engine — Scores the behavioral indicators of emotional regulation capacity: consistency of tone, escalation rate, recovery time after conflict signals, and communication pattern variance.
- Dysregulation Detection Engine — Detects patterns consistent with emotional dysregulation without attaching diagnostic labels. Outputs:
"elevated stress indicators", not"borderline personality disorder". - Regulation-Decision Correlation Engine — Measures the correlation between regulation indicators and decision quality in the same actor over time. High-regulation actors make more consistent decisions.
Code Walkthrough
// Safe emotional regulation scoring
function scoreEmotionalRegulation(signals) {
const indicators = {
toneConsistency: computeToneVariance(signals.messages),
escalationRate: computeEscalationRate(signals.timeline),
recoveryTime: computeRecoveryAfterConflict(signals.conflicts),
patternVariance: computeCommunicationVariance(signals.patterns),
};
const rawScore = weightedAverage(indicators, REGULATION_WEIGHTS);
// Governance: translate internal scores to safe labels BEFORE returning
return governanceWrapper.wrap({
score: rawScore,
label: rawScore < 0.4 ? "elevated-stress-indicators" : "stable-regulation-indicators",
// NEVER: "emotionally_dysregulated", "mental_health_concern", "unstable"
});
}
What to Watch For
- Never output diagnostic language. The governance wrapper enforces this — but do not rely on the wrapper alone. The engine itself should never generate clinical terms.
- Require human review for any score below 0.35. Low regulation scores in legal or employment contexts are high-consequence outputs.