Convergent Validity: Does the Icosa Model Measure What Established Frameworks Measure?

Executive Summary

• Coherence predicts clinical dysfunction with a large effect: the correlation between Coherence and Trap burden (self-reinforcing dysfunction cycles) is rₛ = −.62, R² = .380, meaning 38% of the variance in how many Traps a profile activates is accounted for by a single integration score. This is the dose-response gradient that makes Coherence function like a clinical outcome measure.

• Coherence captures structural information that simpler metrics miss: incremental validity over raw Capacity averages is r = .57, R² = .328, and reporting mean scores alone sacrifices a third of the explainable variance in dysfunction risk. The composite architecture (Gateways, Basins, Fault Lines) earns its complexity.

• Profile types are stable, not random: Formation classifications (the model’s profile shape labels) show large test-retest stability at rₛ = .56, R² = .317. When the same inputs go in, the same structural interpretation comes out. The system behaves as a deterministic measurement instrument.

• Coherence is stable enough to track and sensitive enough to detect change: test-retest stability of r = .48, R² = .230 positions Coherence between rigid trait measures and noisy state measures, exactly where a therapy-tracking metric needs to sit.

• The 20-center grid adds real information beyond aggregate scores: grid-level metrics show r = .46, R² = .216 incremental validity over Domain means, confirming that the relationships among centers carry clinical signal that summaries erase.

• Icosa does NOT replicate the Big Five or HEXACO: convergent mapping yielded rₛ = .01 (negligible, non-significant) between Icosa Domains and established trait dimensions. This isn’t a failure, it’s evidence that the model measures something different from existing instruments.

• The model requires 19 of 20 components to reach the 95% PCA variance threshold, explaining 95.9% of variance — with all 20 centers contributing unique information. It doesn’t collapse into a five-factor structure under principal component analysis. The Icosaglyph is a high-dimensional measurement space, not a relabeled Big Five.

• Across three studies and 30,507 total profile analyses, the evidence converges: Coherence is clinically predictive, temporally reliable, structurally distinct from existing frameworks, and irreducible to simpler summaries. This is a measurement system that earns its complexity.

• The null results are as important as the positive findings. Near-zero correlations with Big Five and HEXACO dimensions confirm that adopting Icosa Atlas adds a new layer of clinical information rather than duplicating what you already have.

• For practice adoption, the combined evidence supports using Coherence as a session-by-session outcome metric, Formation classifications as stable diagnostic anchors, and the full 20-center architecture for treatment planning, with confidence that the system measures something real, distinct, and clinically actionable.

Research Overview

The foundational question for any clinical tool is whether it measures anything real. Not interesting, not novel, but real. Does the central metric behave like a clinical outcome measure? Are the profile classifications stable enough to anchor treatment planning? Does the structural complexity add information, or is it just noise dressed up in geometry?

This research program investigated those questions from three angles across a combined sample of 30,507 synthetic personality profiles processed through the Icosa Atlas engine. The first study tested whether Coherence (the model’s 0-to-100 integration score) predicts clinical dysfunction proxies in a dose-response pattern, the way PHQ-9 scores predict functional impairment. The second examined temporal stability: when similar inputs go in, do similar outputs come out, or does the 20-center architecture amplify noise into chaos? The third mapped the model against established personality frameworks (Big Five, HEXACO, and VIA character strengths) to determine whether Icosa is measuring something these instruments already capture or something new.

The three studies aren’t independent curiosities. They form a single evidentiary chain: clinical predictiveness (does the metric track dysfunction?), measurement reliability (can you trust it across time?), and construct distinctiveness (is it adding information or duplicating what exists?). A tool that’s predictive but unstable is useless for tracking progress. One that’s stable but redundant with the NEO-PI-R doesn’t justify adoption. One that’s distinct but clinically inert is an academic exercise. The Icosa model needs to pass all three tests. The combined evidence shows that it does, with specific strengths and specific constraints that matter for how you’d use it in practice.

Key Findings

Coherence Behaves Like a Clinical Outcome Measure

The most consequential finding across the entire program is the relationship between Coherence and Trap burden. Traps, in the Icosa model, are self-reinforcing feedback loops: Rumination locks Focus x Mental into a cycle that only the Body Gate can break; Codependence Traps Bond x Relational until the Choice Gate opens. They’re the model’s structural analog of clinical symptom cycles. If Coherence is going to function as an outcome metric, it needs to predict how many of these cycles are active in a given profile.

It does. The Spearman correlation between Coherence and active Trap count is rₛ = −.62, p < .001, R² = .380, a large effect accounting for 38% of the variance. This isn’t a threshold effect where “well” profiles have zero Traps and “unwell” profiles have many. It’s a continuous gradient: as Coherence declines from Thriving through Steady, Struggling, Overwhelmed, and into Crisis, Trap accumulation increases in a graded, predictable fashion. Each point of Coherence lost corresponds to measurably increased structural vulnerability.

What makes this finding clinically relevant is the dose-response shape. The PHQ-9 works because higher scores reliably correspond to greater depression severity, not because it explains all the variance in depression, but because the relationship is monotonic and graded. Coherence shows the same property with respect to Trap burden. A client whose Coherence drops from 72 (Steady) to 58 (Struggling) between sessions hasn’t just crossed an arbitrary threshold; they’ve entered a region of the Coherence landscape where more self-reinforcing dysfunction cycles are structurally expected. That’s actionable clinical information.

The dose-response finding also has a boundary worth noting. Coherence’s correlation with Trap severity (how extreme the dysfunction becomes once a Trap activates) was much weaker (r = −.22, R² = .047). Coherence tracks whether you’re trapped, not how deeply. This dissociation between Trap count and Trap intensity suggests that local dynamics (specific Gateway closures, specific Basin configurations) govern severity, while global integration governs susceptibility. Clinically, this means Coherence change tells you the system is moving in the right direction, but you still need to look at individual Trap states to know whether the remaining active Traps are mild or entrenched.

The Composite Metric Earns Its Complexity

A reasonable skeptic might ask: why not just average the 20 center scores and call it a day? The incremental validity findings answer this directly. Coherence predicts clinical outcomes beyond what simple Capacity means provide, at r = .57, p < .001, R² = .328. That’s a third of the variance in dysfunction risk that you’d lose by defaulting to simpler summaries.

Where does this surplus come from? Coherence isn’t a mean. It’s a computed property that incorporates Gateway states (are the nine structurally critical centers open, closed, or overwhelmed?), Basin membership (is the profile locked into a multi-center attractor state like Affective Shutdown or Guarded Scanning?), Fault Line vulnerabilities (where would small perturbations cascade?), and asymmetric penalty weighting (under-expression costs more than over-expression at equivalent distances from center). Two profiles can share identical average center scores yet differ dramatically in Coherence if one has its Gateways open and the other has three closed Gateways feeding active Basins.

The grid-level incremental validity finding from the convergent mapping study reinforces this point from a different angle. The 20-center structure adds r = .46, R² = .216 of information beyond Domain-level means. When you collapse the four Capacities within each Domain into a single number, you lose roughly 22% of the structural signal. The relationships between centers (the pattern of which centers are centered, which are under, which are over, and how those states interact through Gateways and Traps) carry clinical information that no aggregate can capture.

For practice, this means the full Icosaglyph (the model’s 4×5 map of all 20 Harmonies) isn’t decorative complexity. It’s where the treatment-planning information lives. A Domain summary that says “Emotional functioning is moderate” doesn’t tell you whether Empathy is flooded while Discernment is shut down, a configuration that activates the Emotional Flooding Trap and requires the Discernment Gate to resolve. The 20-center architecture captures that distinction. The incremental validity data confirm it matters.

Profile Types Are Stable Measurement Outputs

A personality model that produces different profile classifications every time you run it isn’t a measurement instrument, it’s a random number generator with labels. The temporal stability study tested whether the Icosa model’s outputs hold steady when inputs hold steady, and whether they shift proportionally when inputs shift.

Formation classifications (the model’s 77 profile shape labels, derived from Coherence band and trajectory pattern) showed large stability at rₛ = .56, p < .001, R² = .317. When similar input configurations enter the engine, similar Formation labels come out. This is the kind of reliability that underwrites clinical use: if you profile a client at intake and get a Formation classification of “Frozen” in the Struggling band, that classification reflects the structural input, not scoring noise.

Coherence itself showed medium test-retest stability at r = .48, R² = .230. In classical psychometrics, test-retest correlations below .70 raise reliability concerns. But Coherence isn’t a classical trait score, it’s a computed emergent property of 20 interacting centers, analogous to a network connectivity index. Network-based measures are expected to show moderate rather than high temporal stability because they reflect dynamic system states that fluctuate. A correlation of .48 means Coherence is stable enough to anchor treatment planning (approximately 23% of variance is reliably shared across matched inputs) while remaining sensitive enough to register real configurational shifts. That’s the sweet spot for a therapy-tracking metric: you want it to move when the client moves, not to be so rigid it can’t detect change and not so volatile it registers noise as progress.

One finding adds nuance: center-level cross-profile consistency was negligible (r = .03, R² = .001). Individual center scores don’t hold steady across similar profiles the way global metrics do. This seems alarming until you consider what it means structurally. Two profiles can achieve similar Coherence and similar Formation classifications through different center-level configurations, the same way two people can achieve similar well-being through different psychological routes. The model’s higher-order constructs (Formations, Coherence bands, Basin configurations) extract pattern information that transcends individual data points. Clinical interpretation should operate at the Formation and Coherence level, not at the level of individual center scores in isolation.

The Model Measures Something Established Instruments Don’t

The convergent mapping results clarify this distinctiveness. The convergent mapping study tested whether Icosa Domains correspond to semantically matched dimensions in the Big Five, HEXACO, and VIA frameworks. The Emotional Domain was mapped to Neuroticism/Emotionality. The Relational Domain was mapped to Agreeableness. The expectation was moderate convergent correlations, reflecting shared variance between constructs that seem to describe similar psychological territory.

The result: rₛ = .01, p = .599, R² < .001. Negligible. Non-significant. The Icosa model’s Domain-level scores share essentially zero variance with corresponding Big Five and HEXACO dimensions. Cross-Domain discriminant correlations were equally flat: r = .02, p = .097, R² < .001.

This is the most important null finding in the entire program, and it’s precisely what you want from a new framework. If Icosa Domains had correlated at .50 or .60 with Big Five factors, the model would be measuring substantially the same constructs as the NEO-PI-R. The near-zero correlations confirm structural independence. The Icosa model occupies a different region of personality measurement space.

Why? Because the two frameworks answer fundamentally different questions. Trait models describe what kind of personality someone has: how extraverted, how neurotic, how open. Coherence indexes how well that personality system is organized: whether the centers are coordinated, whether the Gateways are open, whether the system is caught in self-reinforcing dysfunction loops. A person can score within normal limits on every Big Five dimension and still show low Coherence if their Capacity-Domain states are internally conflicted. The Icosa model would flag that through Fault Lines and Basin analysis; the NEO-PI-R would miss it entirely.

The dimensionality analysis drives this home. Principal component analysis of the 20 center scores required 19 of 20 components to reach the 95% variance threshold (95.9% cumulative). The Icosaglyph doesn’t collapse into a five-factor structure. Each of the 20 Harmonies (Sensitivity, Empathy, Curiosity, Intimacy, Surrender, Presence, Discernment, Acuity, Attunement, Vision, Inhabitation, Embrace, Identity, Belonging, Devotion, Vitality, Passion, Agency, Voice, Service) carries unique variance that can’t be reduced to fewer factors; no center is absent or redundant. This is a high-dimensional measurement space, not a rotated or inflated version of existing taxonomies.

Convergent Validity Lives at the System Level

The pattern across all three studies tells a coherent story about where the Icosa model’s validity resides. It’s not at the Domain level, since individual Domains don’t map onto Big Five factors. It’s not at the center level, since individual center scores show negligible cross-profile consistency. Validity lives at the system level: Coherence predicts Trap burden, Formation classifications are stable, and the 20-center architecture adds information beyond what any simpler summary captures.

This is consistent with the model’s theoretical architecture. Coherence isn’t extracted from the centers the way a Big Five factor is extracted from items. It’s computed from the relationships among centers: Gateway states, Basin memberships, Fault Line vulnerabilities, asymmetric penalties. The clinically relevant information is structural, not elemental. A single center score in isolation tells you relatively little. The configuration of all 20 centers, processed through the model’s geometric logic, tells you a great deal.

For adoption decisions, this means the Icosa model’s value proposition isn’t “a better personality test.” It’s “a different kind of personality information.” It complements rather than replaces existing instruments. A practice that already uses the NEO-PI-R for trait assessment gains a structurally distinct layer of analysis by adding Icosa Atlas, one that captures integration, vulnerability patterns, and intervention targets that trait profiles were never designed to reveal.

Boundaries of the Evidence

The null results in this program are not failures; they’re among the most informative findings. The near-zero correlations between Icosa Domains and Big Five/HEXACO dimensions (rₛ = .01 and r = .02) confirm that the model isn’t redundantly measuring what established instruments already capture. If you’re evaluating whether to add Icosa Atlas to a practice that already administers trait inventories, this is the finding that matters most: you won’t be paying for the same information twice. The 19 effective dimensions confirm the same point from a different angle, the Icosaglyph’s measurement space is high-dimensional, not a five-factor model wearing a geometric costume.

The negligible center-level consistency (R² = .001) is equally informative. It tells you that the model’s clinical value doesn’t reside in individual center scores treated as standalone data points. Interpreting a single Harmony score the way you’d interpret a single PHQ-9 item, as a reliable indicator in isolation, would be a misuse. The model’s reliability and predictive power emerge at the level of Coherence, Formations, and structural configurations. This constrains how you’d train clinicians to use the tool: read the Icosaglyph as a pattern, not as 20 independent numbers.

The weak Coherence-severity relationship (R² = .047) sets a practical boundary. Coherence tells you how many dysfunction cycles are active, not how intense each one has become. This means Coherence change is a valid progress indicator for overall structural health, but clinicians still need to examine individual Trap states and Gateway conditions for the granular picture. The model provides both levels of analysis, global Coherence and local Trap/Gateway detail, and the data confirm that both are needed.

Clinical Use

The combined findings reshape how personality assessment fits into clinical workflow. Most practices use trait inventories at intake and symptom measures session-by-session, with a gap between the two: the trait profile tells you who the client is, the PHQ-9 tells you how they’re doing this week, and the connection between structure and symptoms lives entirely in the clinician’s head. Icosa Atlas bridges that gap with a single framework that provides both structural mapping and a trackable outcome metric.

Here’s what the workflow looks like with these findings in hand. At intake, the Standard assessment (32 questions, ~5 minutes) generates the full Icosaglyph: all 20 Harmonies scored, Coherence computed, Gateways assessed, active Traps identified, Basin configurations detected, and a Formation classification assigned. The Coherence score gives you an immediate severity anchor: a client scoring 41 (Overwhelmed band) with seven active Traps is in a structurally different place than one scoring 67 (Steady band) with two active Traps, and the rₛ = −.62 dose-response gradient tells you that difference is clinically meaningful, not arbitrary. The Formation classification (rₛ = .56 stability) gives you a reliable profile type that won’t shift with measurement noise, so you can build a treatment formulation around it.

The Centering Path, the model’s computed intervention trajectory, sequences specific targets based on Gateway states and Basin configurations. Because Coherence captures R² = .328 of variance beyond what Capacity means alone predict, the Centering Path’s prioritization of Gateway unlocking and Basin disruption is grounded in the structural dynamics that actually drive dysfunction risk. If the Choice Gate (Focus × Mental) is closed and feeding 10 active Traps, the Centering Path will target it early. If the Body Gate (Open × Physical) is partially open but the Belonging Gate (Bond × Relational) is overwhelmed, the sequencing reflects that structural dependency. The incremental validity data confirm this isn’t over-engineering, the architectural information predicts outcomes that simpler metrics miss.

Session-by-session, the Timeline feature tracks Coherence change against the baseline expectation that ~23% of variance is stable across matched inputs. A Coherence shift larger than what measurement stability would predict signals genuine structural change, whether improvement (Traps resolving, Gateways opening) or deterioration (new Basins activating, Fault Lines cascading). The temporal stability data give clinicians a principled way to distinguish signal from noise in longitudinal tracking, rather than relying on clinical intuition alone. And because the model’s 20 centers each carry unique variance (19 of 20 components needed to reach the 95% PCA threshold), the information density of each reassessment is high. You’re not just getting a single number moving up or down, but a full structural update across a high-dimensional space.

The safety screening capability, 30 patterns automatically flagged, adds a layer of risk detection that operates independently of clinician judgment. Combined with the dose-response finding that low Coherence reliably predicts high Trap burden, the system provides early warning when a client’s structural profile is deteriorating toward Crisis-band configurations where multiple self-reinforcing dysfunction cycles are simultaneously active.

Applied Example

Consider a 34-year-old client presenting with relationship difficulties and chronic indecisiveness. Standard intake measures (a Big Five inventory, the PHQ-9, the GAD-7) show mildly elevated Neuroticism, moderate anxiety, and subclinical depression. The trait profile says “somewhat neurotic.” The symptom measures say “somewhat anxious.” Neither tells you why this particular person can’t make decisions in relationships or where to intervene first.

The Icosa Atlas Standard assessment reveals a Coherence score of 51 (Struggling band) with a Formation classification that’s been stable across the assessment’s internal consistency checks. The Icosaglyph shows a specific pattern: Agency (Move × Mental) is deeply under-engaged, the client isn’t generating decisional energy. But Acuity (Focus x Mental) is over-engaged, fixating on analysis without resolution. This combination activates the Decisional Paralysis Trap, which requires the Choice Gate (Focus × Mental) to break. The Choice Gate, however, is closed; Acuity’s over-engagement is precisely what’s keeping it shut. Meanwhile, Attunement (Focus × Relational) is over-engaged, activating the Hyperattunement Trap, which requires the Identity Gate (Bond × Mental) to resolve. The Identity Gate is partially open but constrained by an active Detached Surveillance Basin involving under-engaged Embrace and Belonging alongside over-engaged Discernment and Acuity.

Now the clinical picture has structural specificity. The convergent mapping data (rₛ = .01 with Big Five) confirm that this structural information isn’t something the NEO-PI-R was going to give you. “Somewhat neurotic” doesn’t distinguish between someone whose Emotional Domain is flooded and someone whose Mental Domain is locked in an analysis-without-action loop. The incremental validity data (r = .57 beyond Capacity means) confirm that the Trap and Gateway architecture is carrying real predictive weight, and you can’t get this from averaging the client’s four Capacity scores.

The Centering Path prioritizes the Choice Gate first, because opening it would simultaneously provide escape routes for both Decisional Paralysis and the Somatic Obsession Trap that’s also active. The temporal stability data (rₛ = .56 for Formations) give you confidence that this structural formulation will hold across sessions, since you’re not chasing a profile that shifts every week. But the moderate Coherence stability (r = .48) means you should expect the Coherence score itself to respond to genuine therapeutic work. If you successfully help the client reduce Acuity’s over-engagement (perhaps through mindfulness-based interventions that shift Focus from fixating toward attending), the Choice Gate should begin to open, and Coherence should rise as Traps deactivate.

Three sessions later, reassessment shows Coherence has moved from 51 to 57, still in the Struggling band, but the Decisional Paralysis Trap has deactivated. The dose-response data (rₛ = −.62) tell you this Coherence increase corresponds to a meaningful reduction in structural vulnerability. The Hyperattunement Trap is still active, though, and the Centering Path now prioritizes the Identity Gate. The client’s Formation classification hasn’t changed, the temporal stability data predicted this, since Formation-level patterns are more stable than Coherence scores. But within that stable Formation, the internal configuration is shifting. The Clinician Map shows the structural change in detail; the plain-language summary translates it into language the client can use to understand their own progress.

This is what converging evidence from three studies makes possible: not just a number going up, but a structural narrative that explains why it went up, what changed, what hasn’t changed yet, and where to focus next. No single study’s finding produces that clinical capability. The dose-response gradient gives you the outcome metric. The temporal stability gives you confidence in the formulation. The construct distinctiveness confirms you’re seeing something the trait inventory missed. Together, they transform personality assessment from a one-time intake exercise into an ongoing structural guide for treatment.

Connections Across the Research

The findings in this validation family connect directly to results from other study families in the broader Icosa research program. The 19 effective dimensions found in the convergent mapping study independently replicate a finding from the Geometry family’s grid-architecture analysis, which also confirmed that the 4×5 Icosaglyph retains near-full dimensionality rather than collapsing into fewer factors. Two different analytic approaches, applied to different samples, converge on the same structural conclusion, the 20-center architecture isn’t redundant.

The Coherence family’s internal consistency work provides the mechanism behind the dose-response finding reported here. That family’s five-layer Coherence formula demonstrated r = .81 between the computed composite and its structural components, explaining how Coherence integrates Gateway states, Basin dynamics, and penalty functions into a single score. The present validation family’s finding that Coherence predicts Trap burden at rₛ = −.62 shows that this integration works, the composite doesn’t just cohere internally; it predicts clinically relevant outcomes. The States family’s hot-core dynamics (r = .57 between hot-core health and Coherence) explain why the dose-response gradient is so strong: the centers most implicated in active Traps are the same centers that most heavily influence the Coherence score. When hot cores deteriorate, Coherence drops and Traps proliferate, a single structural mechanism producing both the predictor and the outcome.

Operational Impact

The business case for Icosa Atlas adoption rests on three measurable advantages that emerge from this evidence cluster. First, Coherence functions as a structurally interpretable outcome metric with R² = .380 predictive validity for dysfunction burden, comparable to or exceeding the predictive performance of many established clinical screeners. Practices that currently track progress with the PHQ-9 or OQ-45 can add Coherence as a complementary metric that captures structural vulnerability rather than symptom counts, providing earlier warning of deterioration and more specific guidance for intervention targeting. Second, the 19 effective dimensions and near-zero correlation with Big Five instruments mean that Icosa Atlas provides new clinical information rather than duplicating what your existing assessment battery already captures. For practices seeking evidence-based differentiation, this is a concrete claim backed by data: “We assess personality integration and structural vulnerability, not just traits and symptoms.” Third, the Formation stability finding (rₛ = .56) supports using Icosa profiles as anchors for treatment planning across the full course of therapy, reducing the need for repeated full-battery reassessment while maintaining structural specificity.

The efficiency gains are tangible. A Standard assessment takes approximately 5 minutes of client time and produces a Coherence score, Formation classification, Gateway status map, active Trap inventory, Basin configuration, Fault Line analysis, and a computed Centering Path, all automatically generated. The clinical information density per minute of assessment time is high, and the longitudinal tracking capability (Timeline feature) means each subsequent assessment builds on the structural baseline rather than starting from scratch. For group practices and wellness centers managing high caseloads, this translates to more precise treatment planning with less clinician time spent on assessment interpretation.

Conclusion

What this body of evidence establishes is a personality assessment framework with demonstrated clinical predictiveness, measurement reliability, and structural distinctiveness from existing instruments, all three properties confirmed through converging computational analyses. Coherence predicts dysfunction in a graded, dose-response pattern (rₛ = −.62) that supports clinically meaningful severity assessment. Profile classifications demonstrate large test-retest consistency (rₛ = .56) that supports longitudinal treatment planning. And the near-zero convergent correlations with Big Five instruments, combined with 19 effective dimensions, confirm that the model captures structurally distinct information.

For clinical directors evaluating adoption, the practical implication is specific: Icosa Atlas adds a layer of structural personality analysis that complements existing trait inventories and symptom measures, capturing dynamic, relational, Gateway-dependent architecture that those instruments weren’t designed to assess. The dose-response gradient means Coherence change is clinically interpretable: not just “the number went up” but “structural vulnerability decreased in a way that predicts fewer active dysfunction cycles.” The Formation stability means your treatment formulations hold across sessions. The incremental validity means the full 20-center architecture earns its complexity in predictive power.

The overall picture is one of complementarity rather than replacement. The computational evidence demonstrates measurement properties that are distinct from existing instruments and predictive of clinically relevant outcomes. These findings provide a foundation for clinical adoption decisions, with the caveat that empirical validation with human samples remains a necessary next step.

Key Takeaways

• Coherence predicts Trap burden at rₛ = −.62 (R² = .380): treat it as a structurally interpretable outcome metric alongside your existing symptom measures.

• Reporting Capacity averages instead of full Coherence sacrifices R² = .328 of predictive variance; always use the composite score, not simplified summaries, for clinical decision-making.

• Formation classifications show rₛ = .56 test-retest stability; anchor your treatment formulation at the Formation level, where reliability is strongest.

• Icosa Domains share essentially zero variance (rₛ = .01) with Big Five dimensions, the model provides new information that complements rather than duplicates existing trait inventories.

• The 20-center grid uses 19 of 20 components to reach the 95% variance threshold (95.9%), with all 20 centers contributing unique information — the structural complexity is real, not reducible, and clinically informative.

• Coherence stability of r = .48 balances reliability with change sensitivity: expect the score to hold steady when the client is stable and to move when genuine structural change occurs.

• Interpret Coherence change as a structural signal, but examine individual Trap and Gateway states for severity detail, the global metric tracks how many dysfunction cycles are active, while local analysis reveals how entrenched each one is.