Inviting Dissent
Published · Last reviewed
Ask for the objections people actually hold, collect them in writing before anyone speaks, and respond to the first one with a question. Detert and Burris (2007) found that managerial openness was related to employees' improvement-oriented voice, through their perception of how safe speaking up was. Nemeth, Brown and Rogers (2001) found real dissent beat three forms of assigned devil's advocacy on the quantity and quality of solutions.
Durable practice (not new). This is established communication practice, not a discovery or a trend.
What it fixes
The breakdown is the silent yes. You announce a plan, ask “Any objections?”, and hear nothing. The silence feels like agreement, so you proceed, and the concern arrives later as a missed deadline or a quiet exit. Employees judge how risky it is to speak up (Detert and Burris, 2007). In our reading, an open invitation with no structure leaves that judgment where it was, though no study cited here tested that.
A second breakdown is the assigned devil's advocate. A colleague voices objections they do not hold, the group senses the performance, and the discussion tends to confirm the original plan.
How to run it
- Before the discussion, ask each person to write down the strongest concern they actually hold about the plan.
- Open the meeting by naming the decision and what could still change it.
- Invite the written concerns one at a time, starting with the person who usually speaks last.
- Answer the first concern with a question and a thank-you before you offer any defense.
- Close by naming the objections you heard, what you changed because of them or why you kept your course, and where later objections can go.
The written round in step one is a practitioner addition. It gives people time to think and lets objections arrive without the first speaker steering the room. Step three is the same kind of addition: starting with the person who usually speaks last is practitioner consensus, not a tested protocol.
“Thank you. Tell me more about what would have to be true for that concern to go away.”
Say it out loud
“I want the objections you actually hold. If you disagree, I would rather hear it now, and I will tell you tomorrow what I did with it.”
The listener knows exactly what you are doing and why, and the promise to report back gives them a reason to believe it. Never reward dissent in the room and penalize it afterward. If you have already decided, say that, and ask for objections to the way you will carry it out.
Profile translations
- Profile B Before a launch or reorganization decision, send the written-concern request the day before and bring the pile to the meeting with names removed if people prefer.
- Profile C Before you publish a course outline, ask two colleagues which lesson they would cut and why, and wait for an answer you did not expect.
- Profile D Before a pricing change, ask the people who talk to customers every day for the objection they expect to hear first.
- Profile E In a planning meeting with volunteers, ask which part of the plan each person would drop, and write the answers where all can see them.
How to tell it worked
- Action
- Collect written objections before anyone speaks.
- Why
- Writing first lets people state what they hold before the room sets a direction.
- Verification
- The pile includes at least one concern you had not heard before.
- Action
- Answer the first dissent with a question and a thank-you.
- Why
- The first response shows everyone what happens to a person who disagrees.
- Verification
- A second person raises a concern within a few minutes, instead of the room going quiet.
- Action
- Report back within a day on what changed and what did not.
- Why
- Showing what happened to an objection tells people what speaking up costs.
- Verification
- Someone refers to a named objection in the follow-up or adds a new one.
Failure signs: stop if you see these
- Only agreement arrives, from everyone, with no questions.
- You rebut the first concern at length.
- The same person always dissents and is treated as the contrarian.
- Objections are collected and nothing changes, so the next invitation gets silence.
Why it works, and how strong the evidence is
Evidence. Two studies of different kinds. Detert and Burris (2007) is a field study of relationships in one restaurant chain. Nemeth, Brown and Rogers (2001) is a laboratory comparison. Effect sizes are Not reported here, and the written-first step is practitioner consensus without controlled evidence on this page.
Detert and Burris (2007, Academy of Management Journal) studied 3,149 employees and 223 managers in a restaurant chain, in two phases. Managerial openness was more consistently related to improvement-oriented voice than transformational leadership, after controlling for individual differences. The link ran through employees' perceptions of psychological safety, meaning how risky they judged speaking up to be. Leadership behavior had the strongest effect on the voice of the best-performing employees. The abstract describes relationships in one industry, not a randomized test of any technique.
Nemeth, Brown and Rogers (2001, European Journal of Social Psychology) compared an authentic minority with three versions of devil's advocate: the advocate's position known and consistent with the assigned side, known and inconsistent, or unknown. Outcomes were the quantity and quality of solutions. The authentic minority was superior to all three, and the authors point to the difficulty of reproducing authenticity by role-playing. An earlier study by the group had suggested that devil's advocacy fostered thinking aimed at bolstering the initial view.
What is contested or missing: the lab study shows that in that task assigned advocacy did not match real dissent, and says nothing about whether it never helps. Whether the findings carry over to your meetings is untested here.
Worked example
An invented scene, offered as an illustration and not as data. A manager is about to move a team to a new tool.
“That's the plan. Any objections? No? Great. And Sam, play devil's advocate for a minute so we are covered.”
“Yesterday I asked everyone to write down the one thing about this move they would change or worry about. I have the pile here. The one I heard most is that the old tool handles exports and the new one does not. Thank you for writing it. Who can tell me what would have to be true for that to be fine?”
The first version produces a performance of dissent and no information. The second names a real concern, shows what happens to it and invites more. What to watch for: if the same concern has to be raised again next month, people will notice that raising it changed nothing.
Frequently asked questions
Why is assigned devil's advocacy weaker than real dissent?
In Nemeth, Brown and Rogers (2001), an authentic minority produced better results on quantity and quality of solutions than three forms of devil's advocate. The authors note that authenticity is hard to clone by role-playing.
How do I get honest objections from a quiet team?
Ask for written objections before discussion, respond to the first with a question, and report back on what changed. Detert and Burris (2007) linked managerial openness to voice through psychological safety, which suggests that what you do with the first objection matters.
Should I ever play devil's advocate myself?
You can ask a question to test a plan, and say so openly. Presenting a position you do not hold as if it were an objection is the kind of role-play the study found less useful.
What if nobody objects?
Treat silence as information, not agreement. Ask what would lead someone to vote no, or what they would drop, and try the written round next time.
Does this slow decisions down?
It adds a short written round and a response step. The studies cited here did not measure speed, so Not reported.
Want a practice like this each day?
The briefing is published on this site every day. The email edition is not sending yet. Leave your address and you’ll be among the first told when it starts. Nothing is sent in the meantime.