You say you want candour. What actually happens to the person who gives it?

Seven findings from our library that frame the decision.

Voice can cause the retaliation, not just fail to prevent it. An identity threat perspective on employee voice finds that employee voice — constructive, well-intended input — can act as an antecedent of abusive supervision. The pathway runs through leader identity threat: the manager experiences the input as a challenge to their authority or competence, and responds accordingly.

But only for some managers, and the split is stark. In the same study, the indirect effect of voice on abusive supervision was 0.48 for supervisors high in "traditionality" — endorsement of hierarchical role relationships — and −0.09 for those low in it. For low-traditionality supervisors, the association was absent. The risk is not spread evenly across your management population. It is concentrated.

Hierarchy suppresses upward voice specifically. In a study of flight attendants' safety practices, psychological safety positively predicted peer-level and pro-social safety voice, but was negatively associated with upward safety communication. People spoke to each other. They did not speak up the chain. The authors attribute this to hierarchical communication norms.

What we know about voice may be weaker than it looks. An integrative systematic review of employee silence and voice covering 76 studies found 45 distinct measurement instruments and no consensus on what should be measured. Only 10.5% of studies were rated methodologically good; 42.1% were rated poor. Any confident claim in this area, including ours, deserves to be held loosely.

We may also be measuring the wrong thing. The same review argues that safety-specific voice — raising a hazard — likely differs fundamentally from general voice, yet most research conflates the two. The authors suggest this may be why speaking-up interventions have produced such disappointing results.

What does move voice is leader behaviour, not exhortation. Ethical leadership prompts voice through affective commitment — an emotional bond, not a rational calculation. Servant leadership works through employee work reflection, with the mediated path significant (β = 0.228) and substantially amplified among proactive employees, though still present among the least proactive. And supervisor support predicts voice among frontline officers, strengthened by informal relational ties outside the work relationship.

And almost nobody studies the aftermath. A systematic review of abusive supervision found that recovery is a near-absent construct: only two studies examined how subordinates recover from it. The field measures the damage far better than it measures repair.

The evidence does not point one way, and two tensions are worth stating plainly.

The first is that voice appears in this literature as both cause and casualty. It is a behaviour good leadership produces, and a behaviour that can trigger bad leadership. Both can be true, because they run through different managers. That makes "encourage more voice" an incomplete instruction — it depends entirely on who is receiving it.

The second is methodological. A review that rates 42% of its own field's studies as poor is telling you the ground is soft. The identity-threat finding is a single study with a clear mechanism and a clean moderator, which is persuasive but not settled. Treat the direction as credible and the magnitude as provisional.

Yes, speaking up gets punished — but not randomly, and not everywhere. It gets punished where a manager experiences input as a threat to their standing. That is a property of the manager, not of the employee's courage or the company's stated values.

This reframes the problem. Most voice initiatives act on employees: reassure them, give them a channel, tell them the door is open. The evidence suggests the more useful target is the small number of managers for whom subordinate input registers as insubordination. If you have those managers, no channel will help, because the people reporting to them are making an accurate risk assessment.

It also means silence is a signal about your management, not about your people.

Ask who was last punished for being right. Not whether people feel safe — that is a survey question and it will be answered politely. Ask a specific one: name the last person who raised something uncomfortable, and describe what happened to them afterwards. If nobody can name one, that is your answer.

Look at voice by manager, not by company. Any aggregate score hides the concentration the research describes. If your engagement data can be cut by reporting line, cut it, and look for teams that are quieter than their neighbours.

Separate hazard reporting from opinion. The evidence suggests these are different behaviours with different drivers. If you run one channel for both, you are likely measuring neither well.

Stop treating an open door as an intervention. It offers access. The research points to leader behaviour — ethical conduct, support, genuine attention to input — as what actually changes whether people use it.

The mechanism in this research is hierarchy: whether a manager holds a strongly hierarchical view of the manager–subordinate relationship, and therefore reads upward input as a challenge. That construct travels well to Gulf workplaces, where steep hierarchy, seniority norms and informal relational networks all shape who can say what to whom.

But be careful with that inference, because we cannot yet support it with local evidence. The traditionality studies here were conducted in Chinese and Confucian settings, and the supervisor-support study measures guanxi — a specific form of informal tie. Nothing in our library tests this in a GCC organisation. Treat the hierarchy mechanism as a well-founded hypothesis to check inside your own company, not as a finding about the Gulf.

That check is cheap. If the research is right, the quiet teams in your organisation will cluster under particular managers rather than spreading evenly. You can look at that this quarter without commissioning anything.

All seven sources below are open-access under CC BY 4.0.

1. An identity threat perspective on why and when employee voice brings abusive supervision — Frontiers in Psychology
2. An integrative systematic review of employee silence and voice in healthcare: what are we really measuring? — Frontiers in Psychiatry
3. Abusive Supervision: A Systematic Review and New Research Approaches — Frontiers in Communication
4. The effects of psychological safety and employee voice behavior on flight attendants' mindful safety practices adoption — Frontiers in Public Health
5. How Ethical Leadership Prompts Employees' Voice Behavior: The Roles of Affective Commitment and Moral Disengagement — Frontiers in Psychology
6. Servant Leadership and Employee Voice Behavior: The Role of Work Reflection and Proactive Personality — Frontiers in Psychology
7. When and why voice to higher-up? Declaring the psychological mechanisms of subordinate's voice behavior in the public sector — PLOS ONE