The SIGNAL™ framework. A definitive guide for 2026.
Five layers. Six nodes. Twenty-one sub-clusters. Sixty-three behavioral proxies. The first framework that decomposes psychological safety into weighted, observable signal — not survey opinion.
Bytru.place
Published4 Jun 2026
Read14 min
AudienceManager
Amy Edmondson defined psychological safety in 1999 and gave the field a measurement instrument: the seven-item Psychological Safety Scale. That scale is still the academic gold standard. It is also a self-report — and self-report has a ceiling that operational diagnostics need to break through. SIGNAL™ is the framework tru.place built to operationalise the construct: behavioral signal in place of self-report, continuous in place of annual, weighted by predictive power rather than equal-weighted by survey item. This article walks through every layer.
What is the SIGNAL™ framework?
SIGNAL™ is the tru.place Intelligence System for behavioral measurement of psychological safety. It is the architecture underneath every score, every alert, every diagnosis the product produces. The framework has four design choices that distinguish it from the survey-first instruments most companies use.
Behavior over opinion. The unit of measurement is observable action — speak-up frequency, recognition timing, decision autonomy exercised — not Likert-scale self-report. Behavior is harder to fake and anchored to specific events.
Continuous over annual. Signals are captured in 30-second channel-native check-ins, daily. Trajectory matters as much as level: a score of 61 today means nothing; a 61 that was 74 three weeks ago means everything.
Weighted over equal. The six nodes are not equally predictive. Safety to Speak and Authenticity Gap carry 20% each; the remaining four carry 15%. The weights reflect predictive power in cascade and exit-risk modeling.
Diagnostic over descriptive. A score map alone is too coarse to act on. SIGNAL™ Diagnose names the pattern, attributes the cause, prescribes the surgical action, and predicts the impact.
The five layers
Each layer is a different lens on the same underlying signal.
SIGNAL™ Core — the measurement layer. The 6-node score map. Daily check-ins, 30 seconds, channel-native. This is what every team sees first.
SIGNAL™ Pulse — the trajectory layer. Score in 4 weeks, three scenarios (no action, partial action, full echo). Score without trajectory is a thermometer reading; trajectory is the temperature trend.
SIGNAL™ Echo — the reciprocity layer. Manager logs an action, team is notified anonymously that something changed because of signals like theirs. Without echo, data degrades within weeks; with echo, the flywheel runs.
SIGNAL™ Equity — the DEI intelligence layer. Same team score of 68 can hide a 26-point gap across demographic groups. Equity makes the gap auditable under CSRD ESRS S1 and Pay Transparency Directive obligations.
SIGNAL™ Diagnose — the surgical layer. 24 documented patterns (leader-caused, team-caused, system-caused, DEI-caused). Each pattern carries a trigger signature, a root-cause attribution, a one-action prescription, and a predicted impact.
The six nodes
The six nodes are the operational core of the framework — the layer most managers see daily. Each node represents a category of observable behavior that, when present, indicates psychological safety; when absent, indicates risk.
S
Safety to Speak
20% · highest
I
Initiative Behavior
15%
G
Group Belonging
15%
N
Need for Autonomy
15%
A
Authenticity Gap
20% · highest
L
Learning Comfort
15%
Source: tru.place SIGNAL™ framework · S and A weighted highest as most predictive of exit risk†
Each node maps to a specific behavioral signature.
Safety to Speak (S) — 20%. Do people raise concerns in meetings, push back on decisions, admit when something is not working? Speak-up frequency in meetings, dissent in code review, willingness to say "I think we are wrong about this." S is weighted highest because it is the first node to drop in the cascade — the earliest leading indicator.
Initiative Behavior (I) — 15%. Do people propose ideas without being asked? Voluntary contribution, proactive scope expansion, suggestions in 1:1s before the manager prompts. I stalls one to two weeks after S drops.
Group Belonging (G) — 15%. Do members signal inclusion in informal moments — meetings, threads, recognition? Slack thread inclusion, post-meeting catch-ups, who gets the recognition emoji. G erodes before formal exclusion shows up in performance data.
Need for Autonomy (N) — 15%. Do people exercise judgment, or wait for permission? Decision authority used, method autonomy in execution, willingness to deviate from prescribed approach. The Deci & Ryan self-determination construct, operationalised.
Authenticity Gap (A) — 20%. Is there a measurable gap between the version of the person at work and the person at home? Emotional labour signals, identity disclosure, the distance between professional persona and private self. A is the strongest single predictor of 90-day exit risk†.
Learning Comfort (L) — 15%. Do people admit not knowing, ask for help, ask basic questions? Admitted-uncertainty frequency, help requests, "I don't know" in postmortems. L is the last node to drop in the cascade and the first to recover when safety is rebuilt.
Why are the nodes weighted?
Because they are not equally predictive. Most psychological-safety instruments treat survey items as equal weights — a sum or average across the items, no calibration for which item is most diagnostic. SIGNAL™ inherits the academic construct but calibrates the weights against three lines of evidence.
First, cascade order. Internal pattern modeling shows S is consistently the first node to drop when psychological safety begins to break, and A is the strongest predictor at the 30-to-90-day exit horizon. Both deserve higher weight in any early-warning system.
Second, Edmondson's own scale construction. The 1999 PS Scale has seven items; three of them map directly to what SIGNAL™ separates as S, and one maps to A. The original instrument over-weights speak-up safety implicitly through item count — SIGNAL™ makes that weight explicit.
Third, Aristotle's findings. Project Aristotle identified five team dynamics; the two that emerged as most predictive — psychological safety and dependability — both anchor in what is observable as speak-up behavior and follow-through on commitments. S and I together carry most of that signal.
"Psychological safety was far and away the most important of the five key dynamics we found — it's the underpinning of the other four."
Google re:Work · Project Aristotle · 2015
The 90-day cascade
The framework's predictive power comes from understanding that the six nodes do not move independently. They cascade — sequentially, reasonably consistently across teams.
Week 0
S
People stop raising concerns
Week 1–2
I
Initiative stalls
Week 2–3
A
Authenticity gap widens
Week 3–4
L
Learning shuts down
Week 6–8
exit
Flight risk realised
The SIGNAL™ cascade — S → I → A → L → exit · 30-60 day intervention window before the consequences crystallise†
The cascade is the failure mode. It is also the intervention window. A manager who knows the cascade is forming has 30 to 60 days of optionality before the visible consequences arrive. A manager who only sees the annual engagement number has, on average, none.
When Authenticity Gap exceeds 30 points on a 0–100 scale, our modeling indicates a 67% probability of exit within 90 days†. The number is projected — see footnote. The cascade pattern itself, however, is well-documented in the academic literature. Morrison and Milliken's 2000 work on Organizational Silence describes how silence becomes a collective property of a team, even when each individual would, alone, choose to speak.
How SIGNAL™ differs from Edmondson's scale
Edmondson's seven-item Psychological Safety Scale, introduced in her 1999 paper Psychological Safety and Learning Behavior in Work Teams, is the most widely used self-report measure in organisational research and the academic foundation of the field. SIGNAL™ inherits the construct and operationalises it differently. The four design differences:
Unit of measurement. Edmondson: self-report Likert items on a 1–7 scale. SIGNAL™: behavioral proxies on a 0–100 normalised scale, drawn from observable action.
Frequency. Edmondson: typically annual or once-per-study. SIGNAL™: daily 30-second check-ins, with the unit of analysis being the rolling four-week trajectory.
Decomposition. Edmondson: seven items aggregated into one team-level score. SIGNAL™: 63 behavioral proxies aggregated into 21 sub-clusters and 6 weighted nodes.
Output. Edmondson: a research-grade construct measurement. SIGNAL™: an operational diagnostic — pattern named, cause attributed, action prescribed, impact predicted.
Both have a place. For academic research at the team level, Edmondson's scale remains the gold standard — it is validated, replicable, and comparable across decades of published research. For operational diagnosis in a working team, where the manager needs to know what is happening this week and what to do about it, SIGNAL™ is the framework built for that work.
The deeper reason SIGNAL™ measures behavior rather than opinion is one borrowed from Japanese: tatemae. The face one shows in public. Honne is what one actually thinks.
Workplaces in every culture have their own version of tatemae. The polite answer in a one-on-one. The slightly upbeat response on Slack. The careful phrasing in a channel where the manager is reading. Most surveys measure tatemae and call it data.
"What people show at work is not what people think at work. Most HR dashboards measure the show."
tru.place
The closest observable proxy for honne is behavior. Did the person speak up in the meeting? Did they ask the question that risked looking uninformed? Did they admit a mistake in front of the team? Did they push back on a senior colleague's plan? These behaviors aggregate over weeks into a picture that describes the team more honestly than any single survey can. SIGNAL™ is the framework that lets that aggregation happen systematically.
Where SIGNAL™ is today, and where it is going
The MVP product surfaces SIGNAL™ Core: the 6-node score map, daily check-ins, trajectory across rolling four-week windows, basic threshold alerts. Manager dashboard at Enterprise tier, individual score for Collaborator tier.
The 2026 roadmap brings the remaining four layers online progressively. SIGNAL™ Diagnose (24 documented patterns with root-cause attribution and surgical actions) is the core differentiator versus engagement-dashboard products. SIGNAL™ Pulse (4-week prediction with three scenarios) turns measurement into forecast. SIGNAL™ Echo (closed-loop reciprocity) is the flywheel fuel that prevents response decay. SIGNAL™ Equity (intersectional DEI signal) is the audit-grade layer for CSRD ESRS S1 and Pay Transparency Directive obligations.
The framework is described in this article in its full architecture because the public product roadmap is public, and the design choices made today shape what an organisation can act on in three months and what it cannot. Teams that adopt SIGNAL™ at the Core layer today gain the trajectory dataset that Diagnose, Pulse, Echo, and Equity will build on.
Pre-launch · 50 teams per month
The framework Amy Edmondson proved. The diagnostic engine tru.place built.
tru.place is launching in 2026. Join the waitlist at tru.place — we're letting in 50 teams a month. Risk intelligence for people decisions. Built in Europe. Designed for the world.
SIGNAL™ is the tru.place Intelligence System for measuring psychological safety through behavioral signals. It decomposes the construct into 5 layers (Core, Pulse, Echo, Equity, Diagnose), 6 nodes (Safety to Speak, Initiative Behavior, Group Belonging, Need for Autonomy, Authenticity Gap, Learning Comfort), 21 sub-clusters, and 63 behavioral proxies.
Why are the nodes weighted?
The nodes are weighted because they are not equally predictive. Safety to Speak is weighted 20% because it is the first node to drop in the cascade — the earliest leading indicator. Authenticity Gap is also weighted 20% because it is the strongest single predictor of 90-day exit risk. The remaining four nodes carry 15% each.
How is SIGNAL™ different from the Edmondson scale?
Amy Edmondson's seven-item Psychological Safety Scale is a self-report instrument designed for academic team-level research. SIGNAL™ is a behavioral-proxy framework designed for operational diagnosis: it measures what people do rather than what they say, runs continuously rather than annually, and produces actionable trajectory data rather than a single point-in-time score.
What does SIGNAL™ Diagnose do?
SIGNAL™ Diagnose is the layer that turns the 6-node score map into a named pattern (Pattern L1 Blame Culture, Pattern T3 Authenticity Crisis, and so on across 24 documented patterns), attributes the root cause (leader-caused, team-caused, system-caused, or DEI-caused), prescribes a surgical action this week, and predicts the impact.
Is SIGNAL™ ready to use today?
SIGNAL™ Core (the 6-node score) is in the MVP. SIGNAL™ Diagnose, SIGNAL™ Pulse, SIGNAL™ Echo, and SIGNAL™ Equity are on the 2026 roadmap. The framework is described in this article as the diagnostic engine tru.place is building; the public product reflects the MVP scope at any given launch milestone.
Sources & notes
Edmondson, A. (1999). Psychological Safety and Learning Behavior in Work Teams. Administrative Science Quarterly, 44(2), 350–383. DOI: 10.2307/2666999
Edmondson, A. (2018). The Fearless Organization. Wiley.
Google re:Work (2015). Guide: Understand team effectiveness — Project Aristotle. Retrieved 2026-05-22 from rework.withgoogle.com.
Morrison, E. W. & Milliken, F. J. (2000). Organizational Silence: A Barrier to Change and Development in a Pluralistic World. Academy of Management Review, 25(4), 706–725.
Deci, E. L. & Ryan, R. M. (2000). Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being. American Psychologist, 55(1), 68–78.
† Projected from internal tru.place pattern modeling. The 67% exit probability at A > 30, the cascade station timing, and the 30–60 day intervention window are unvalidated internal figures. tru.place will publish validated retention probability and cascade-window data after six months of operational data, in line with our commitment not to misrepresent unvalidated figures.