Psychological Safety Scale (PSS-Team)

Abstract

The Psychological Safety Scale (PSS-Team), conceptualized and validated by organizational behavior scholar Amy C. Edmondson in her landmark 1999 investigation, stands as the premier psychometric instrument for assessing team-level psychological safety within organizational, clinical, educational, and frontline service environments. The instrument measures the shared tacit belief held by members of a work unit that the team environment is secure for interpersonal risk-taking—specifically, that individuals will not be embarrassed, rejected, marginalized, or penalized for speaking up with questions, concerns, unconventional ideas, or candid reports of mistakes. Structurally, the scale comprises 7 self-report items evaluated along a 7-point Likert scale ranging from 1 (“Very inaccurate / Strongly disagree”) to 7 (“Very accurate / Strongly agree”), featuring three negatively phrased, reverse-scored items designed to mitigate acquiescence response bias.

Psychometrically, the PSS-Team has demonstrated robust cross-cultural and cross-industry reliability and construct validity. In Edmondson's foundational study across 51 diverse work teams in an office design manufacturing corporation, the scale exhibited high internal consistency (Cronbach's alpha = .82) and substantial within-unit agreement (mean $r_{wg}$ exceeding .80), justifying its aggregation from individual perceptual responses to a shared, emergent team-level emergent construct. Subsequent confirmatory factor analyses and structural equation models across thousands of independent work units—ranging from intensive care unit (ICU) surgical teams to software engineering squads (e.g., Google's Project Aristotle)—have affirmed its unidimensional structural stability, discriminant validity from related constructs like team cohesion and trust, and profound predictive validity regarding collective learning behavior, voice behavior, employee retention, innovation, and error-detection efficacy.

Keywords

psychological safety, team learning behavior, Amy Edmondson, interpersonal risk-taking, organizational psychology, team dynamics, error management, speaking up, voice behavior, psychometrics

Authors

The Psychological Safety Scale was developed and established by:

• Amy C. Edmondson, Ph.D.: Novartis Professor of Leadership and Management at Harvard Business School, Boston, Massachusetts, United States. Dr. Edmondson completed her doctoral studies in Organizational Behavior at Harvard University under the mentorship of foundational theorists including Chris Argyris and J. Richard Hackman. Her research centers on human interactions, collaborative teaming, psychological safety, and organizational learning across high-reliability organizations, healthcare settings, and modern knowledge-work sectors.

Inquiries regarding academic theoretical frameworks and research applications can be directed to Harvard Business School, Soldiers Field, Boston, MA 02163, USA.

Purpose

The primary purpose of the Psychological Safety Scale is to quantify an elusive yet foundational collective psychological state: the tacit, taken-for-granted climate regarding the interpersonal consequences of personal vulnerability within work groups. While historical organizational literature often collapsed group climate into broad dimensions such as mutual trust, group cohesion, or satisfaction, Edmondson recognized that teams frequently cultivate warm social rapport while concurrently harboring intense anxiety regarding the exposure of errors, the admission of ignorance, or the questioning of entrenched managerial orthodoxies. The PSS-Team was intentionally constructed to bridge individual cognitive assessments of risk with collective organizational outcomes, providing a rigorous diagnostic mechanism to identify why some groups continuously adapt, innovate, and catch catastrophic errors early, whereas structurally identical groups experience catastrophic oversights and organizational silence.

In industrial and organizational settings, the clinical diagnostic value of the scale lies in identifying latent vulnerabilities in communication pipelines. In high-stakes environments—such as aviation flight decks, hospital operating theaters, nuclear facilities, and financial audit departments—the absence of psychological safety does not merely suppress employee morale; it directly increases preventable morbidity, accidents, and compliance breaches. When team members calculate that asking a clarifying question will brand them as incompetent, or reporting a minor near-miss will trigger disciplinary vengeance, they systematically prioritize immediate personal image management over collective organizational welfare. The PSS-Team isolates these defensive calculations, transforming subjective dread into an empirically trackable metric.

From an academic and organizational research perspective, the PSS-Team serves as a pivotal explanatory variable, operating frequently as a primary mediator or moderator in multi-level structural equation frameworks. Researchers utilize the instrument to model how leadership styles—specifically inclusive, transformational, and servant leadership—translate into objective operational performance. By measuring team psychological safety, scholars can evaluate how diversity initiatives yield intellectual synergy rather than intergroup friction, how agile workflows influence team reflexivity, and how frontline service units buffer burnout while proactively executing customer recovery protocols. In consumer and frontline service research, the scale explains how team interpersonal environments foster an outward-facing service climate, enabling frontline representatives to deviate from rigid scripts, solve novel customer grievances, and engage in constructive voice without fear of managerial retribution.

Psychological Construct

Team psychological safety is formally delineated as a shared belief held by members of a team that the team is safe for interpersonal risk-taking. To comprehend this psychological construct in depth, it is vital to contrast it with conceptually proximate yet distinct phenomena, examine its multi-level dynamic emergence, and dissect the behavioral manifestations captured by the seven-item architecture of the scale.

Differentiating Psychological Safety from Trust and Cohesion

A frequent theoretical error involves treating psychological safety as a synonym for generalized trust or group cohesion. As Edmondson (1999) argued, psychological safety diverges from interpersonal trust across several structural lines:

• Locus of Focus: Trust typically describes an individual’s expectation concerning another specific actor (e.g., “I trust my supervisor to keep their word”). In contrast, psychological safety describes a generalized climate directed toward the collective group environment (e.g., “How will the team react if I bring forward an unexpected error?”).
• Temporal Horizon: Trust involves conferring benefit of the doubt regarding future potential actions (giving others the power to harm or help down the road). Psychological safety is immediate and situational; it pertains to whether the actor experiences an imminent threat to their professional identity, self-worth, or social standing right now upon taking an interpersonal risk.
• Social Cohesion: Team cohesion involves mutual liking, group attractiveness, and social bonding. Paradoxically, highly cohesive groups often demonstrate severely compromised psychological safety due to groupthink, wherein members suppress divergent viewpoints, critiques, or uncomfortable data to preserve pleasant social harmony. Psychological safety, conversely, encourages productive task conflict, direct dissent, and candid confrontation without imperiling mutual respect.

Core Dimensions Captured by the Construct

While the PSS-Team operates empirically as a parsimonious unidimensional scale, it integrates three interrelated psychological dynamics:

• Non-Punitive Error Orientation: Reflected in items assessing whether mistakes are held against team members. In a psychologically safe environment, an error is conceptualized not as an indictment of individual competence, but as vital informational feedback regarding systemic fragility, training gaps, or complex environmental variability.
• Openness to Difference and Divergence: Evaluated through items examining whether individuals are rejected for being distinct, and whether their unique skills, unconventional perspectives, and idiosyncratic capabilities are valued. Psychological safety eliminates pressures toward homogenizing conformity, allowing neurodiverse, interdisciplinary, and minority perspectives to challenge legacy processes.
• Calculated Reciprocal Vulnerability: Captured by items regarding the ease of seeking help, the willingness to discuss difficult dilemmas, and the conviction that no colleague will deliberately sabotage or undermine another's efforts. Asking for assistance requires admitting a boundary in one's knowledge; psychological safety neutralizes the fear that such transparency will be weaponized in political turf battles or performance appraisals.

Theoretical Framework

The genesis of the Psychological Safety Scale rests upon intersecting streams of organizational sociology, human relations theory, organizational learning models, and cognitive appraisals of workplace threat. The construct synthesizes foundational insights from Kurt Lewin, Edgar Schein, Warren Bennis, Chris Argyris, and Donald Schön.

The Lewinian and Schein Foundation: Unfreezing and Anxiety

The lineage of psychological safety traces back to Lewin's three-stage change model (unfreezing, moving, refreezing) and its subsequent expansion by Edgar Schein and Warren Bennis (1965) in their work on organizational change. Schein identified that for individuals or social systems to adopt new cognitive frameworks, behaviors, or operational technologies, they must experience “unfreezing.” However, unfreezing inherently induces two competing psychological forces:

1. Survival Anxiety: The realization that failing to adapt or learn threatens one's ongoing survival, relevance, or career within the organization.
2. Learning Anxiety: The defensive terror that attempting to learn, asking basic questions, exposing one's lack of mastery, or failing during early experimentation will result in public humiliation, loss of status, or formal sanction.

Schein posited that leaders cannot successfully compel organizational change by merely escalating survival anxiety, as this triggers chronic paralysis and defensive cognitive rigidity. Instead, the only sustainable mechanism is to lower learning anxiety by establishing a climate of psychological safety. Edmondson extended Schein's individual-level change thesis into the realm of modern knowledge-intensive work teams, demonstrating that psychological safety is not a luxury or a comfort mechanism, but a functional imperative for survival in complex, volatile, and uncertain task environments.

Argyris, Defensive Routines, and Double-Loop Learning

A second pillar of the theoretical framework derives from Chris Argyris and Donald Schön's (1978) conceptualization of organizational defensive routines and double-loop learning. Argyris articulated that human beings operate according to deeply internalized “theories-in-use” that emphasize self-protection, winning at all costs, and suppressing negative affect (termed “Model I” behavior). In typical corporate environments, when errors arise, teams engage only in single-loop learning: they adjust tactical steps to correct the symptom while rigidly preserving governing assumptions, policies, and systemic norms. Double-loop learning—which demands questioning underlying norms, challenging leadership dogmas, and acknowledging fundamental strategic miscalculations—is routinely avoided because it exposes individuals to defensive vulnerability.

Edmondson embedded this insight directly into the PSS-Team architecture. Team learning behavior (e.g., experimentation, collective reflection, boundary spanning, and questioning assumptions) is costly in terms of interpersonal risk. Without a pervasive, group-level emergent property that reassures members that vulnerability will not be exploited, individuals default instinctively to impression management and organizational silence. Consequently, the PSS-Team operationalizes the climate preconditions necessary to dismantle Argyrisian defensive routines, allowing genuine double-loop learning to occur within everyday project teams.

Validity

The construct, convergent, discriminant, and criterion-related predictive validity of the Psychological Safety Scale has been subjected to rigorous empirical evaluation across more than two decades of global organizational research.

Construct and Structural Validity

In her foundational 1999 study, Edmondson deployed the scale across 51 teams encompassing 427 employees within a large manufacturing corporation specializing in office furniture. Construct validity was tested by contrasting psychological safety with theoretically distinct team dimensions, including team efficacy (Bandura's social cognitive framework) and team cohesion. Confirmatory analysis confirmed that while efficacy measured the group's self-belief in its operational capability to accomplish tasks, psychological safety uniquely indexed the interpersonal consequences of risk. Multitrait-multimethod correlation matrices confirmed that psychological safety scores shared moderate, healthy correlations with supportive leadership ($r = .35$ to $.55$) while retaining clean, divergent structural loading.

Criterion and Predictive Validity

The predictive validity of the PSS-Team is extraordinary, demonstrating robust associations with both perceptual behavioral outcomes and hard, objective operational milestones:

• Team Learning Behavior: In the initial validation study, structural equation models demonstrated that psychological safety exerted a profound direct effect on team learning behavior ($eta = .53, p < .001$), acting as the essential mediating mechanism between team structure, supportive leadership, and collaborative performance.
• Objective Team Performance: Internal customer satisfaction metrics and executive performance ratings correlated significantly with team psychological safety, operating through the transmission pathway of team learning ($R^2 = .46$ for performance outcomes explained by the full structural model).
• Healthcare Safety and Near-Miss Detection: In longitudinal clinical trials within hospitals (e.g., Edmondson, 1996; Nembhard & Edmondson, 2006), higher scores on the PSS-Team initially appeared counter-intuitively correlated with higher detected medication error rates. Further investigation confirmed that teams with high psychological safety were not making more mistakes; rather, their members felt safe to acknowledge, log, and discuss near-misses and errors, whereas low-safety wards systematically concealed mistakes until clinical catastrophes occurred.
• Google's Project Aristotle: In a comprehensive multi-year internal study examining 180 active engineering and product teams at Google, researchers analyzed dozens of quantitative variables—from team composition, IQ, and seniority to social dynamics. The findings, published in 2015, identified psychological safety (measured via adaptations of Edmondson's 7-item instrument) as the single most critical structural bedrock distinguishing hyper-performing teams from underperforming counterparts. Teams scoring high on the PSS-Team exhibited significantly lower turnover, leveraged diverse ideas more effectively, and consistently exceeded revenue targets.

Cross-Cultural and Trans-Disciplinary Generalizability

The PSS-Team has undergone successful psychometric validation across diverse linguistic and cultural landscapes, including European, East Asian, and Latin American contexts. Studies conducted in high-power-distance cultures (e.g., South Korea, Japan, China) demonstrate that while the baseline mean of psychological safety may be shifted by cultural deference to hierarchical authority, the scale's factor structure remains invariant, and its predictive leverage on innovation and error-reporting remains universally vital.

Reliability

The internal consistency, inter-rater reliability, and stability coefficients of the Psychological Safety Scale consistently exceed standard psychometric benchmarks established by the American Psychological Association (APA).

Internal Consistency Reliability

In Edmondson's (1999) initial validation cohort, the 7-item scale produced a Cronbach's alpha of .82, demonstrating high internal consistency without redundant item overlap. Across subsequent meta-analyses synthesizing hundreds of team-level studies (e.g., Frazier et al., 2017), the internal consistency of the scale has reliably oscillated between $lpha = .78$ and $lpha = .91$. The deletion of any single item from the 7-item set generally fails to elevate the omnibus alpha, confirming the harmonious conceptual contribution of both the forward-scored and reverse-scored items.

Aggregation Statistics and Within-Group Agreement

Because the PSS-Team is designed primarily as a team-level construct measured via individual reports, psychometric justification for aggregating individual scores to the group level requires rigorous verification of within-group consensus and between-group variance. Psychometric evaluations consistently report:

• $r_{wg}$ Index of Within-Group Agreement: James, Demaree, and Wolf's $r_{wg}$ coefficient frequently exhibits median and mean values exceeding $.80$ (with Edmondson reporting an initial mean of $.84$), well above the conventional $.70$ heuristic required to justify data aggregation.
• Intraclass Correlation Coefficients (ICC): Studies systematically report statistically significant ICC(1) values—representing the proportion of variance attributable to team membership—ranging between $.15$ and $.38$ ($p < .001$). ICC(2) values—indexing the reliability of the team means—routinely surpass $.70$, establishing that the aggregated team score provides a dependable, stable assessment of the group's shared interpersonal climate.

Test-Retest Stability

In longitudinal organizational field experiments, test-retest reliability across 3- to 6-month intervals has demonstrated moderate to high stability ($r = .65$ to $.78$). The scale captures an enduring environmental climate rather than a fleeting affective mood, while concurrently retaining sufficient sensitivity to detect genuine changes in team climate triggered by targeted leadership interventions, organizational restructuring, or psychological safety training programs.

Factor Analysis

The latent structural integrity of the Psychological Safety Scale has been thoroughly evaluated through both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Structure

During scale construction, Edmondson subjected a broad pool of candidate items tapping risk-taking, error orientation, help-seeking, and interpersonal respect to principal components analysis with varimax and oblimin rotations. The items loaded cleanly onto a single primary factor accounting for the majority of the common variance, with eigenvalues exceeding conventional scree test cutoffs. Items with ambiguous cross-loadings or weak communalities were pruned, yielding the definitive, highly focused 7-item instrument.

Confirmatory Factor Analysis and Model Fit

Subsequent confirmatory factor analyses in diverse organizational samples have repeatedly established the superiority of the unidimensional single-factor model over alternative multi-factor specifications. Structural equation modeling metrics consistently meet or exceed contemporary criteria for adequate model fit:

• Comparative Fit Index (CFI): Values routinely range from $.95$ to $.99$.
• Tucker-Lewis Index (TLI): Values consistently exceed $.93$.
• Root Mean Square Error of Approximation (RMSEA): Estimates typically fall between $.03$ and $.06$, with 90% confidence intervals well below the $.08$ threshold for acceptable fit.
• Standardized Root Mean Square Residual (SRMR): Values consistently remain beneath $.05$.

Factor Loadings of Items

Standardized factor loadings ($lambda$) across published literature demonstrate robust item-to-construct linkages. In Edmondson's data and subsequent major structural investigations, standardized loadings for all seven items generally fall between $.55$ and $.84$:

• Item 1 (mistakes held against you – reverse scored): $lambda pprox .68$ to $.76$
• Item 2 (bring up problems and tough issues): $lambda pprox .70$ to $.82$
• Item 3 (reject others for being different – reverse scored): $lambda pprox .58$ to $.72$
• Item 4 (safe to take a risk): $lambda pprox .75$ to $.84$ (often the highest loading anchor item)
• Item 5 (difficult to ask for help – reverse scored): $lambda pprox .62$ to $.74$
• Item 6 (undermines efforts): $lambda pprox .64$ to $.78$
• Item 7 (unique skills valued and utilized): $lambda pprox .60$ to $.75$

Methodological research testing whether the three reverse-scored items form an artifactual “method factor” shows that while incorporating an orthogonal method factor for negatively worded items can slightly improve statistical indices, the substantive interpretation of the underlying single psychological safety construct remains exceptionally robust and unaltered.

Instrument / Measurement Tool

• Test Type: Multi-item psychometric self-report questionnaire designed for aggregation to the group, team, or departmental level of analysis.
• Target Population: Members of functional, cross-functional, clinical, project, or operational work teams within organizational, educational, and service environments.
• Item Count: 7 items.
• Response Format: 7-point Likert scale (1 = Very inaccurate / Strongly disagree to 7 = Very accurate / Strongly agree).
• Reverse-Scoring Protocol:
  • Items 1, 3, and 5 are negatively keyed and must be reverse-scored prior to calculating scale metrics.
  • Mathematical conversion formula for 7-point scale: $\text{New Score} = 8 – \text{Raw Score}$.
• Scoring and Aggregation Procedure:
  • Individual Scale Score: After reverse-scoring Items 1, 3, and 5, calculate the arithmetic mean across all 7 items for each respondent. Mean scores range from 1.0 to 7.0, where higher scores reflect elevated psychological safety.
  • Team-Level Aggregation: Verify within-team agreement ($r_{wg} ge .70$) and intraclass correlations (ICC1 and ICC2 with statistical significance). Compute the collective mean of the individual respondent scores across all verified members of that intact work unit.
• Administration Time: Approximately 3 to 5 minutes to complete, minimizing survey fatigue when administered as a standalone diagnostic or embedded within broader employee climate assessments.

Permissions & Fee and Test Year

The Psychological Safety Scale (PSS-Team) was published in 1999 within the peer-reviewed scholarly article authored by Dr. Amy C. Edmondson in Administrative Science Quarterly. In accord with standard academic conventions, the instrument items as published in academic literature are widely accessible and freely available for non-commercial, academic, scientific, and educational research purposes, provided that appropriate scholarly attribution is formally cited.

For large-scale proprietary commercial use, corporate-wide consulting interventions, or integration into fee-based software platforms, permissions may be subject to the copyright policies of the publisher (SAGE Publications / Administrative Science Quarterly) or formal institutional licensing protocols associated with Harvard Business School. Researchers seeking official guidance or validation certifications can refer to Dr. Edmondson's publicly available organizational and academic publications.

References

• Argyris, C., & Schön, D. A. (1978). Organizational learning: A theory of action perspective. Addison-Wesley. https://eric.ed.gov/?id=ED160844
• Edmondson, A. (1996). Learning from mistakes is easier said than done: Group and organizational influences on the detection and correction of human error. The Journal of Applied Behavioral Science, 32(1), 5–28. https://doi.org/10.1177/0021886396321001
• Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383. https://doi.org/10.2307/2666999
• Edmondson, A. C. (2018). The fearless organization: Creating psychological safety in the workplace for learning, innovation, and growth. John Wiley & Sons. https://www.wiley.com/en-us/The+Fearless+Organization-p-9781119477242
• Frazier, M. L., Fainshmidt, S., Klinger, R. L., Peachey, A., & Rathert, C. (2017). Psychological safety: A meta-analytic review and extension. Personnel Psychology, 70(1), 113–165. https://doi.org/10.1111/peps.12183
• Nembhard, I. M., & Edmondson, A. C. (2006). Making it safe: The effects of leader inclusiveness and professional status on psychological safety and improvement efforts in health care teams. Journal of Organizational Behavior, 27(7), 941–966. https://doi.org/10.1002/job.413
• Schein, E. H., & Bennis, W. G. (1965). Personal and organizational change through group methods: The laboratory approach. John Wiley & Sons. https://catalog.hathitrust.org/Record/001099616

Items of the Scale