Why staff surveys and organisational metrics can give leaders false comfort

“In regulated industries cultural measurement is not optional. Bosses and boards are expected to monitor opinions and demonstrate oversight of behavioural and conduct risk. Yet surveys can mislead if interpreted incorrectly. And those who treat engagement levels as reflections of organisational health risk confusing statistical stability with behavioural soundness. Odessa S. Hamilton argues that this “illusion of cultural health” arises when favourable majorities are mistaken for organisational strength. In staff surveys taken at one global financial group, a trading division consistently scored above the organisational average, with completion rates often exceeding 70 per cent. But when misconduct surfaced, internal reviews found that several staff had expressed concerns, while assuming their reservations were not widely shared. No one wanted to appear disloyal and that reticence had been mistaken for contentment. The warning signs were lost in averages. This dynamic is not unusual. In the years before crises became public, several major financial institutions could point to internal metrics that suggested cultural soundness. Engagement appeared strong; cultural scores steady; escalation metrics reassuring; with e thical climate indicators meeting benchmarks. These results flowed upward to the board, where oversight mechanisms were in place. Still, serious conduct failures unfolded beneath the surface. In high-profile enforcement cases, politicians and regulators later reported that governance structures, compliance functions and reporting processes were active as misconduct persisted . This paradox is not accidental. In high-stakes institutions, where poor decisions, misconduct, or operational failures can have serious consequences for stakeholders, cultural and behavioural risk measurement does not occur in a vacuum; it unfolds under regulatory scrutiny, incentive pressure, reputational risk and deep internal loyalties. Surveys can be thoughtfully designed yet still mislead if interpreted incorrectly. They are sometimes poorly designed which further muddies the waters. But more often they are naively interpreted. Either way, leaders who treat engagement levels and favourable majorities as direct reflections of organisational health risk confusing statistical stability with behavioural soundness and silence with strength. Regulation shaping the signal In financial services and other regulated industries, cultural measurement is no longer optional. Leaders are expected to evidence monitoring and boards must demonstrate oversight of behavioural and conduct risk, with regulators reviewing survey outputs as part of supervisory dialogue. This pressure has consequences. Completion rates often become performance indicators in their own right, with arbitrary response thresholds set without methodical justification. A 60 per cent participation target is routinely treated as a marker of credibility and regulatory readiness, despite the statistical basis for that figure being seldom examined. Thresholds implicitly suggest that deriving results would be sufficiently representative to support reliable conclusions. But raw participation is not shorthand for reliability, and reliability does not guarantee validity. Whether results can be trusted depends on context. And findings should not be judged by the presence of a symbolic majority, but whether respondents differ meaningfully from non-respondents, whether measures capture what they intend to capture and whether responses are free from material bias. When leaders are held to engagement targets and evaluated on response rates within their units, a subtle, unintentional behavioural shift emerges that conceals the very risks the survey is meant to surface. Firms can mandate participation, but they cannot mandate transparency and “ carrot or stick” strategies lead to compliance without true engagement, where straight-lining increases, with reduced variation between answers. (Straight-lining is a response pattern in which participants repeatedly choose the same rating across multiple items, often used as an indicator of reduced response quality.) A dataset can appear robust because the denominator is large, but a pressured 75 per cent completion rate may be less informative than a voluntary 45 per cent rate. When measurement becomes a target Regulatory expectations can also shape how results are framed. When results are destined for review, there is a tendency to highlight strengths, emphasise improvements and contextualise anomalies. It reduces appetite for confronting uncomfortable outcomes, particularly if they threaten the narrative of control. Regulatory pressure can unintentionally incentivise tidy metrics, which echoes a well-established scientific principle: when a measure becomes a target it loses its diagnostic value . In organisational settings, results matter, as they influence performance evaluations, reputational standing and senior conversations. So people become motivated to lean to neutrality or Likert highs, meaning outliers tail off and alignment increases on paper. (Likert highs refer to responses at the upper end of a rating scale (typically “Agree” and “Strongly Agree”) that are used as indicators of favourable sentiment.) The paradox is the more serious cultural metrics are taken in governance frameworks, the more likely they are to be behaviourally shaped. The meaning of silence While high response rates should not be mistaken for organisational health, low rates should prompt deeper inquiry. Participation below 20 per cent presents another interpretive challenge and is a datapoint in itself. Not all explanations are nefarious; they may simply reflect operational overload during peak reporting, audit or deal cycles. Timing matters and survey windows should be appropriately scheduled, with results interpreted with seasonal priorities in mind. It may result from listening saturation, where overlapping assurance exercises, such as conduct attestations, behavioural risk surveys and compliance certifications, leave employees fatigued. Assigned time for completion and co-ordination across functions can reduce redundancy, while improving result quality. Low engagement may also signal “learned futility” after repeated surveys fail to produce visible change, eroding confidence that contribution matters. Visible follow-through and feedback loops are critical. It may reflect “quiet quitting” where employees reduce discretionary effort while remaining employed. These patterns should be interpreted alongside retention, absenteeism and mobility data rather than in isolation. More concerning is a distrust of anonymity or where loyalty norms discourage dissent. Here, external assessors, anonymity assurances, and qualitative triangulation become critical. When most employees decline to respond, that silence is not without value, it is data requiring contextualisation. The averaging problem Even with broad participation, aggregation conceals risk. Most reporting mechanisms emphasise mean scores, but organisational fragility often resides in concentration, not central tendency. A subculture representing 10 to 15 per cent of a workforce can sustain unhealthy norms while barely shifting the organisational average. In risk management, leaders understand concentration risk intuitively, but culture requires the same logic. Monitoring lower-tail concentration is essential. If more than 10 per cent of respondents select “disagree” or worse, “strongly disagree” on core questions, it warrants scrutiny regardless of headline scores. Concentrated exposure, not average compliance, determines risk and cultural indicators require the same lens. Tracking variance and distributional shape over time often reveals emerging pockets of risk earlier than mean trends. Equally, sudden score uniformity can be alarming, which is why repeated measurement with a consistent, well-defined instrument and a strong interpretive framework is important. Identity, loyalty, and psychological safety Organisational identification can motivate behaviours that appear to benefit the organisation at the expense of rules. Individuals who strongly identify with their institution may experience criticism as self-threatening, and loyalty can suppress voice. Psychological safety can moderate this effect . But where it is low, silence may be rational rather than apathetic. Pluralistic ignorance compounds the risk, where individuals privately doubt norms but assume others accept them, and aggregated positive results reinforces this assumption. The result what might be called “the illusion of cultural health”, where stable scores, high completion rates and limited variance masking localised fragility. The response cannot be only encouraging people to “speak up”. Leaders must distinguish loyalty to mission from hierarchical loyalty, ensuring that challenge is framed as contribution rather than disloyalty. Structural protections with independent escalation channels, visible protection of dissenters and public examples where constructive challenge altered decisions, matter more than slogans. Together they help to recalibrate a perceived cost of voice. When employees see that dissent is neither isolated nor punished, silence becomes less rational. Psychological safety can be engineered through consistent managerial behaviour and institutional design. Without these counterweights, even well-designed surveys will continue to measure compliance more reliably than sincerity. Leaders overseeing cultural metrics should routinely make three assessments First, identify where the concentrations of dissent are, even if the average appears healthy. Second, explore non-respondents and assess whether they systematically differ from those who responded. Third, from a behavioural perspective, evaluate whether external factors may have shaped responses. Without disciplined interrogation, surveys risk providing assurance rather than insight. Surveys remain highly valuable tools and pair well with interviews, focus groups, desktop reviews, and ethnographic methods to surface behavioural risk, ethical grey areas, informal norms and early signs of misconduct. They offer breadth, pattern detection, relational insight and trend visibility. The aspiration is to exploit survey data appropriately and assess results more meticulously. “Interpretative maturity” means moving beyond headline scores to examining distributions, tracking participation, evaluating missingness, interrogating trends over time and scrutinising sudden shifts. Results must be contextualised against geopolitical events, HR metrics, workload cycles and other organisational influences, while exercising caution when linking participation to short-term reward or punishment. Remember that measurement is never neutral; incentives shape expression, loyalty constrains voice, identity reframes criticism and regulatory pressure shapes reporting. Leaders who recognise these forces are less likely to be comforted by pacifying dashboards and more likely to detect emerging risks while they are still contained, avoiding public crises. The illusion of cultural health is rarely preceded by a lack of data. More often it arises when favourable majorities are mistaken for organisational strength. This article gives the views of the author, not the position of LSE Business Review or the London School of Economics. You are agreeing with our comment policy when you leave a comment. Image credit: Jack_the_sparow provided by Shutterstock. The post Why staff surveys and organisational metrics can give leaders false comfort first appeared on LSE Business Review .

Summary generated from the RSS feed of LSE Business Review. All article rights belong to the original publisher. Click through to read the full piece on blogs.lse.ac.uk/businessreview.