The elephant is the room
Introduction
Ask any mediator whether the setting matters and you will get a yes. Book a decent room, arrange the chairs sensibly, put water on the table. Ask what else the room does and the conversation dries up. The space is treated as logistics: something to arrange before the real work of interests, emotions and process design begins.
That treatment is hard to defend. Decades of research in environmental psychology show that physical spaces systematically shape stress, attention, cooperation and decision making. In other words, the room is not the backdrop of the conversation, but a participant in it.
Let’s translate that to a mediation: it is 15:40, the second straight hour in a small meeting room with the door closed. The air has gone stale, everyone is slightly too warm. The party across the table seems more stubborn than they did this morning. The mediator seems a little less sharp, maybe a little less impartial. Nobody in that room thinks: the room is doing this. The frustration lands on the people, because people are visible and air is not.
That is the precise sense in which the mediation room is not neutral: it pushes the conversation in one direction or another, whether we want it or not, and whether we notice it or not.
What this essay does
Five things:
It shows how the room works on the bodies and stress systems that parties bring in with them.
It examines the geometry of the room: distance, tables and sightlines.
It looks at ownership: whose space it is, on and off the screen, and what that does to power.
It covers fuel and timing, the least glamorous levers with some of the best evidence.
And it separates what research factually supports from the folklore that circulates in negotiation training.
Each section ends with low-hanging fruit: small interventions that cost little, need no special budget or permission, and can be explained openly to parties.
A word on the evidence, because this essay leans on studies and not all studies are equal. In writing it, I sorted them into three kinds. Some findings have been repeated by independent researchers; those I treat as solid. Some were retested and the effect never showed again; those I name for what they are, debunked. And many exist as a single study that nobody ever ran twice; those I use, but I say so in the same sentence, and I never let them carry more weight than one study can bear.
Where I could, I traced findings back to their original sources; results tend to grow in the retelling, and this field retells a great deal. And thin evidence was allowed to support only advice that passes a second test: it must be worth following on its ordinary merits, with the research as encouragement rather than foundation. Because the purpose here is not certainty. It is to hand you variables: things in the room that deserve a conscious look where before there was none.
The body in the room
parties bring their bodies and stress systems into the room, not just files and mandates
stale air, heat and fatigue get misread as stubbornness or bad faith
practical cue: irritation that rises with the clock rather than with the content
how to work with this: treat breaks as maintenance for the room, not just for the people
Parties do not walk into a mediation as rational decision makers. They bring bodies that respond to their surroundings long before any argument is exchanged. And of the responses that walk in with their physical bodies, the most studied one involves air: a Harvard team led by Joseph Allen had office workers spend days in environments with different air quality and measured their cognitive performance. In a simulated green office with low chemical emissions, scores were 61 percent higher than in a conventional setup, and 101 percent higher when doubled ventilation was added on top. The extra ventilation lifted scores by another quarter.
An earlier experiment by Usha Satish and colleagues found the same pattern for decision making specifically: at CO2 levels around 1,000 ppm, scores on a decision-making test battery dropped by up to a quarter. At 2,500 ppm, by roughly half.
For calibration: outdoor air sits around 420 ppm. Below roughly 800, thinking stays sharp. An ordinary meeting room with the door closed can pass 1,000 within the hour. If you want to see this happen rather than take it on faith, a consumer CO2 meter costs about fifty euros and fits in a laptop bag.
Researchers argue about the exact thresholds, and about how much CO2 itself contributes; both studies were small, and they share a test battery and an author. But the broader finding survives: stale, poorly ventilated rooms dull thinking. Brains in stale air get worse at precisely what mediation demands: weighing options, taking another perspective, holding back the sharp remark. Readers of Kahneman will recognise the pattern: slow, deliberate System 2 thinking fades first; and fast, reflexive System 1 takes over.
Temperature follows the same logic with smaller effects: office performance declines by roughly one to two percent per degree above 25 degrees Celsius. And noise and echo make complex verbal work harder, a finding mostly from classrooms that no adult meeting room contradicts.
Low-hanging fruit
Book the room half an hour early and air it. You do not want to start at the 1,200 ppm the previous meeting left behind. In a sealed building, ask the venue to boost the ventilation instead.
Break every hour, and let the break serve the room: door open, window open, and, if possible, people out.
When irritation rises with the clock rather than with the content, call a break. Better still, blame the room out loud, truthfully: the air in here has gone heavy, let’s open things up. Side catch: blaming the room hands both parties a face-saving way back.
Geometry: distance, tables and a shared wall
people who compete sit opposite each other; people who cooperate sit side by side, and the table announces which of the two is expected
table shape is symbolism, seat position is science
practical cue: parties addressing you instead of each other
how to work with this: give them something to face together
In the 1960s, Robert Sommer studied where people choose to sit. Pairs expecting to compete took seats opposite each other. Pairs expecting to cooperate sat side by side. Pairs who just wanted to talk chose the corner of the table. This does not prove that opposite seats cause conflict. It proves that opposite seating is the posture people associate with competing, which means a long table with lawyers massed on either side announces the format before anyone has said a word.
Distance does similar subtle work. Edward Hall mapped the zones people reserve for different kinds of contact: roughly up to 1.2 metres for personal conversation, beyond that for social and formal dealings. A three-metre boardroom table parks everyone in the formal zone for the entire day and makes a personal word feel out of place.
The table itself is a different story. Research has plenty on seats and almost nothing on tables. Head seats are real: in classic jury studies, the people seated at the short ends of a rectangular table talked more, were rated as more influential and were chosen as foreman far more often.
The round table’s reputation for equality, on the other hand, is symbolism rather than science. The idea goes back to the twelfth-century Arthur legend, and the rare head-to-head experiment that compared round with rectangular tables found no difference. Even in a circle, people mostly respond to the person sitting opposite them, at least in groups without a leader.
That does not make shape meaningless. It makes it symbolic, and symbols are exactly what parties in conflict read. In the winter of 1968 to 1969, the delegations at the Vietnam peace talks in Paris spent ten weeks of a shooting war negotiating the shape of the table. A triangular table was vetoed immediately for what it implied about sides. The eventual solution was an unmarked round table with two small rectangular tables placed opposite each other: deliberately ambiguous, so each side could read its own geometry into it. When President Johnson left office, the shape of the table was roughly all that had been agreed.
The shape Paris vetoed is the one I do wonder about for mediation: a triangle, two parties and the conflict itself on the third side. But there is no research on it. Still, a flipchart builds that triangle for free, and lets you test it in your next session. Facilitate the conversation through it and the dynamic shifts: two parties that were facing each other are now facing the same wall. The problem gets its own side of the room.
Another idea worth trying is rearranging the room mid-session. It looks like logistics, but it can be a process intervention if you want it. Inviting parties to move to the flipchart often loosens a conversation that got stuck in fixed positions.
Myself, I often take another chair at the foot of the table, or simply stand up and lean against the wall or next to the flipchart. Either way the chart takes my spot at the head of the table.
The most radical layout has no table at all. In Marily Oppezzo and Daniel Schwartz’s experiments, walking lifted the generation of ideas for roughly four in five participants, and slightly hindered the focused thinking needed to weigh them. Conflict researchers have added the geometric argument: walking puts people side by side, moving at a matched pace, with no eye lines to defend. In other words: walk to generate options, sit down to choose between them.
I do realise that this is not for every conflict or every pair of parties: both must be able to walk it comfortably, and the route must be private. But in a deadlock, when the table has stopped producing anything, it is a creative option with real evidence behind the walking part.
Low-hanging fruit
Where you sit is a signal. The head seat presides, the opposite seat confronts, the corner is where conversations happen: the natural choice for individual intakes. In the joint session, choose your seat knowing parties read it.
Have the flipchart standing in the room before parties arrive, and write on it whatever the parties need to face together: the agenda, the timeline, the question of the moment. It is also a great way to pick up naturally on what has been discussed in a previous session.
When the table stops producing, propose a walk. Decide back at the table.
Ownership: whose room is it
negotiating in your own office measurably improves your outcomes
neutrality is shown, not declared
practical cue: one party moves like a host, the other like a visitor
how to work with this: a venue nobody owns, or strictly equal rituals when that fails
Graham Brown and Markus Baer ran three studies on what negotiators call home advantage and found it is real: occupants of a space out-negotiate visitors to it. The measured mechanism is confidence: hosts negotiate with more of it, visitors with less, and an intervention that boosted the visitors’ confidence shrank the gap.
Does the same hold for the mediator? Nobody has measured it: the studies tested negotiators, not neutrals. But confidence in a space is not negotiator-specific; it travels. For a mediator the case may be even stronger.
Running a mediation is running two jobs at once: the process and the logistics. In an unfamiliar venue the second job never stops. Will the projector work, is the second room actually free, when is the coffee coming, the water, lunch; and who do I call if the heating fails? In your own space those questions were answered before the parties arrived. You know where everything is, you know the facilities meet your standard, you know the front desk has your back if something breaks. All the attention those logistics claim is attention you cannot give to the parties and their conflict.
And that ease is not private; parties read it. A mediator who is visibly at home in the room signals, without saying a word, that the process is in hands that know what they are doing. In heavier commercial cases especially, with dominant characters who are used to owning whatever room they walk into, this can make the difference between chairing the process and chasing it.
Leading is partly presence, and presence comes easier where you are the host. Which puts the mediator’s own office in an interesting position: it is home turf, but yours rather than theirs. Foreign to both parties in equal measure, and home to the person guarding the process. One condition: tell parties why you propose your office (the neutrality and the facilities), and offer them an alternative. The advantage stays legitimate only when it is named; choosing remains the parties’ job. Party autonomy is underrated, including when room decisions need to be made.
There is also one seat worth singling out: the one facing the door. Hosts tend to take it without thinking, which leaves visitors with their backs to the room. The preference itself is documented: people have chosen seats with a view of the room since researchers first watched them do it in cafés in the 1960s. Whether a back to the door measurably costs attention has never been tested. But based on the seating research, it is safe to say it costs comfort and control, which affects attention indirectly anyway. In my experience, feeling it as a party made it real, and I pay attention to it now as a mediator: where the room allows it, both parties get an equal view of the door.
Ownership also operates at a smaller scale, and this is where mediators get audited without knowing it. Parties read neutrality from the parking lot onward: who gets greeted first, who gets the chair facing the window, who sits with their back to the door, whose side of the table the good light falls on. None of this is imagined. People under threat scan for status signals, and a mediation is a room full of people under threat. Assume they keep score of things you do not notice. Aside from the plenary room, most mediators also have at least one caucus room. If the two caucus rooms are unequal, rotate the parties between them, and sweep each room before the swap: a strategy sheet left on the table is a breach you organised.
Also, do not leave acoustic privacy implicit. Actual privacy is the duty, and perceived privacy is what unlocks disclosure on top of it. People who can hear the corridor assume the corridor hears them, and hold back exactly the disclosure the caucus exists to make possible. Check the sound insulation beforehand, then tell the parties what you checked: I sat in there earlier and heard nothing from the corridor. It is one more reason a neutral venue beats even a good room at the company’s offices: the venue can be perfectly soundproof and still not feel private.
The window is worth assigning on purpose too. Looking at greenery or a long view lowers heart rate and other stress markers, and it restores the capacity for directed attention, the exact capacity that a day of careful listening drains. There is something simpler at work too: a view gives people somewhere to look besides the table and each other. So if one room has the view, give it to the caucus: that is where the flooded party goes to recover.
The walls themselves deserve a thought as well. What we know about art comes mostly from hospitals: calm, representational nature imagery tends to lower stress in anxious people, while abstract or ambiguous art can do the opposite; in one study, heart-surgery patients facing abstract images fared worse than those facing a blank wall. Nobody has studied paintings in negotiation rooms. But mediation rooms hold anxious people, and the inference is cheap: hang the calm landscape, and if the only art available is edgy, prefer the bare wall.
Low-hanging fruit
When the stakes justify it, book a space neither party owns.
Make the welcome visibly even-handed: same greeting, same walk to the room, the same offer of coffee, equal chairs.
Check the sound insulation before the session and tell parties you did. People only use privacy they can feel.
The room with the window should preferably be the caucus room.
Fuel and timing
hunger makes hard conversations harder
sharing a meal builds agreement; sharing a task does not
practical cue: the afternoon session where every proposal suddenly hardens
how to work with this: no one should decide anything important hungry
Brad Bushman followed 107 married couples for three weeks, measuring evening blood glucose alongside a nightly aggression measure that sounds odd and works well: each evening, participants could push up to 51 pins into a doll representing their spouse, a standard research proxy for aggressive impulses. The lower the evening glucose, the more pins, night after night. The mechanism is debated and the study has its critics. But the practical bet is easy and cheap: hungry people fight harder (and every parent knows it), and a mediator asking two tired parties for hard decisions at 17:00 is working against it.
Food can also build, not just protect. In Lakshmi Balachandra’s negotiation experiments, pairs who shared a meal created 11 to 12 percent more joint value, and her control condition showed the effect came from eating together, not from doing something together. A small study, but it converges with an old hospitality ritual: same food, same table.
The small talk around food and coffee is not filler either. In one experiment, strangers who had five minutes of personal phone chat before a week of email negotiation built far more rapport and deadlocked less often, though that second result was statistically thin. One study, moderate evidence, but it points the same way as the meal: the social wrapper does work the substance cannot.
And then the famous judge study: Israeli judges who granted parole generously right after their breaks and almost never right before them. It gets quoted in every negotiation training, and it is shakier than its fame. Later analyses showed the cases were not randomly ordered, and separate work argued the effect was simply too large to be believable. So skip the neuroscience and keep the habit. Schedule a break before every decisive moment. If the effect is real, you have protected the decision. If it is not, you have lost ten minutes.
And the pause pays twice. What we know about lasting settlements is that they follow from process, not pressure: in a follow-up study of mediated agreements, compliance months later tracked how fair the process had felt and whether everything had been said, not how good the deal looked on paper. I know colleagues who deliberately squeeze towards the end, since tired parties concede faster. Those moves do the process no favour if a lasting decision matters more than a quick one. That is my experience, and the compliance research points the same way. My advice: when the moment comes, party autonomy is the final card: lay out both options with their consequences. Push through now and agree it stays closed, or break for half an hour and decide for good. Then let them choose.
Low-hanging fruit
Do not schedule the hard decisions at the end of a long stretch or just before lunch.
One shared platter, same food for everyone, eaten at the same table. Not individual orders behind laptops. Ask about dietary needs beforehand; same food means equally accommodated, not identical.
The screen is a room too
online, nobody has home advantage: everyone is home
what the mediator works with is a chain of partial rooms, mostly out of sight
practical cue: eyes drifting to something off-screen
how to work with this: name the invisible rooms in the ground rules
Moving online does not escape the room, but multiplies it instead. Jeremy Bailenson’s work on video fatigue names the loads: a face at conversational close-up for hours, the constant sight of yourself, a body pinned to a chair to stay in frame. And for the mediator there is a deeper loss, one I can best describe as mediating under water, or with bubble wrap around your head. Everything you read in a physical room, posture, hesitation, what hangs in the air between two people, arrives compressed into a small tile with a half-second delay.
And in one way no room could be more neutral: online, nobody has home advantage, because everyone is home. Each party sits in a space of their own choosing, set up to their own comfort. That is a genuine benefit, and it explains much of the enthusiasm. But it comes at a structural price: there is no mediation room anymore, only a meeting link and the platform behind it.
Teams, Zoom and Google are buildings in their own right, each with its own doors and corridors: chats, notes, recording buttons, breakout rooms. What the mediator works with is a chain of partial rooms. Who sits just out of frame? Can the colleague in the next cubicle follow along? Is the connection secure, where do shared files end up, and what is being noted on a second device? None of these questions has a visible answer, and every one of them is the kind of doubt that makes people hold back the disclosure mediation runs on.
So the honest framing is a trade, not a verdict. On one side, the accumulated comfort of everyone in their own space. On the other, a mediator who can read less and inspect nothing. And the mediator’s own tech skill becomes part of the room: someone who can open a breakout room without breaking stride is the online version of the host who knows where the light switches are. Fumbling with the settings is the new fumbling with the thermostat.
Whether that trade is worth it depends on the case and the parties. What does not depend on the case: the invisible rooms must be named. Ground rules, agreed out loud, are the online equivalent of checking the sound insulation yourself.
Low-hanging fruit
Open with ground rules for the virtual room: everyone alone in the room, no second devices, no recording, headphones on.
Five-minute setup check with each party beforehand: camera at eye level, face lit, equal framing, and confirm who else can hear them.
Self-view divides people: some steer better when they can see how they come across, others are only distracted by it. Make sure everyone knows hiding it is an option, then leave it alone.
Shorter blocks online: forty-five minutes, then a real screen break. Stand up, stretch, leave the room. The break is the window you cannot open.
On stagecraft
the best-known room tricks are the worst supported
some findings failed the retest; most never got one
practical cue: any wisdom that promises to reprogram a counterpart with a prop
Negotiation lore loves stagecraft: the warm cup, the hard chair, the high ceiling, the clean scent, the circle of chairs. Each carries the same promise: reprogram your counterpart with a prop. The promise does not survive contact with the evidence, which comes in two kinds: the ones that failed when someone looked again, and the ones nobody ever looked at twice.
The first category has a clear champion: you have probably heard that a warm cup of coffee in someone’s hands makes them warmer towards you and your proposal. When three labs reran the warmth-priming work behind it with proper samples, the effect was gone.
The second category is larger. Hard chairs supposedly make hard negotiators. High ceilings supposedly make people think bigger. Same for clean scents and fairness, bright light and heightened emotion, circular seating and a sense of belonging. Each rests on a single paper that nobody ever managed to confirm a second time. None of this is necessarily wrong. It is just confirmed by nobody, and for a mediator that amounts to one instruction: hold it loosely.
Should you stop serving coffee, then? No. Serve it because hospitality opens people up, a ritual as old as negotiation itself. The warm mug never made the difference; the small talk it invites does. Choose comfortable chairs because fatigue is real and four hours is long, and because every discomfort nagging at someone is attention that is not going to the conversation. Dim the harsh lights for glare and fatigue alone. Keep the rituals, drop the certainty. And above all, make party autonomy the reflex: check in, and let the parties decide.
Conclusion
None of this means that a good table layout fixes a bad process. It does not, and the evidence never claimed it did. The claim is more modest and more useful: small contextual choices decide whether a difficult conversation stays just bearable enough to be had in earnest, or becomes colder and more strained than its content required.
The five areas are checkable in ten minutes:
Air: can it be refreshed, and is there a break rhythm that refreshes it?
Geometry: is there something to face together, and room to change the setup mid-session?
Ownership: who owns this space, and is the welcome even-handed?
Fuel: are people fed, and is the hard part scheduled when they are?
Screen: if online, are the invisible rooms named in the ground rules?
What the room changes stays invisible to the people it changes. Parties will keep attributing the heaviness of the afternoon to each other, because people are visible and air is not. That misreading is not going away; they have other things on their minds. Someone in the room should know better. It might as well be the person who booked it.
References
Allen, J.G., MacNaughton, P., Satish, U., Santanam, S., Vallarino, J., & Spengler, J.D. (2016). Associations of cognitive function scores with carbon dioxide, ventilation, and volatile organic compound exposures in office workers: A controlled exposure study of green and conventional office environments. Environmental Health Perspectives, 124(6), 805-812.
Satish, U., Mendell, M.J., Shekhar, K., Hotchi, T., Sullivan, D., Streufert, S., & Fisk, W.J. (2012). Is CO2 an indoor pollutant? Direct effects of low-to-moderate CO2 concentrations on human decision-making performance. Environmental Health Perspectives, 120(12), 1671-1677.
Seppänen, O., Fisk, W.J., & Lei, Q.H. (2006). Effect of temperature on task performance in office environment. Lawrence Berkeley National Laboratory, LBNL-60946.
Klatte, M., Bergström, K., & Lachmann, T. (2013). Does noise affect learning? A short review on noise effects on cognitive performance in children. Frontiers in Psychology, 4, 578.
Xu, A.J., & Labroo, A.A. (2014). Incandescent affect: Turning on the hot emotional system with bright light. Journal of Consumer Psychology, 24(2), 207-216.
Sommer, R. (1965). Further studies of small group ecology. Sociometry, 28(4), 337-348; and Sommer, R. (1969). Personal Space: The Behavioral Basis of Design. Prentice-Hall.
Hall, E.T. (1966). The Hidden Dimension. Doubleday.
Dosen, A.S., & Ostwald, M.J. (2016). Evidence for prospect-refuge theory: A meta-analysis of the findings of environmental preference research. City, Territory and Architecture, 3, 4.
Ulrich, R.S. (1991). Effects of interior design on wellness: Theory and recent scientific research. Journal of Health Care Interior Design, 3, 97-109; and Ulrich, R.S., Lundén, O., & Eltinge, J.L. (1993). Effects of exposure to nature and abstract pictures on patients recovering from heart surgery. Psychophysiology, 30(Suppl. 1), 7 (conference abstract).
Morris, M., Nadler, J., Kurtzberg, T., & Thompson, L. (2002). Schmooze or lose: Social friction and lubrication in e-mail negotiations. Group Dynamics: Theory, Research, and Practice, 6(1), 89-100.
Pruitt, D.G., Peirce, R.S., McGillicuddy, N.B., Welton, G.L., & Castrianno, L.M. (1993). Long-term success in mediation. Law and Human Behavior, 17(3), 313-330.
Strodtbeck, F.L., & Hook, L.H. (1961). The social dimensions of a twelve-man jury table. Sociometry, 24(4), 397-415.
Steinzor, B. (1950). The spatial factor in face-to-face discussion groups. Journal of Abnormal and Social Psychology, 45(3), 552-555.
Minton, H.L., Steger, J.A., & Smrtic, G.R. (1968). Group opinion change as a function of circular vs rectangular seating arrangement. Psychonomic Science, 12(8), 357-358.
Michaels, J. (2017). Stuck in endless preliminaries: Vietnam and the battle of the Paris peace table, November 1968 to January 1969. Defence-in-Depth, King’s College London.
Brown, G., & Baer, M. (2011). Location in negotiation: Is there a home field advantage? Organizational Behavior and Human Decision Processes, 114(2), 190-200.
Oppezzo, M., & Schwartz, D.L. (2014). Give your ideas some legs: The positive effect of walking on creative thinking. Journal of Experimental Psychology: Learning, Memory, and Cognition, 40(4), 1142-1152.
Webb, C.E., Rossignac-Milon, M., & Higgins, E.T. (2017). Stepping forward together: Could walking facilitate interpersonal conflict resolution? American Psychologist, 72(4), 374-385.
Bushman, B.J., DeWall, C.N., Pond, R.S., & Hanus, M.D. (2014). Low glucose relates to greater aggression in married couples. PNAS, 111(17), 6254-6257.
Balachandra, L. (2013). Should you eat while you negotiate? Harvard Business Review (hbr.org digital article), January 2013.
Yin, J., Yuan, J., Arfaei, N., Catalano, P.J., Allen, J.G., & Spengler, J.D. (2020). Effects of biophilic indoor environment on stress and anxiety recovery: A between-subjects experiment in virtual reality. Environment International, 136, 105427.
Kaplan, S. (1995). The restorative benefits of nature: Toward an integrative framework. Journal of Environmental Psychology, 15(3), 169-182.
Bailenson, J.N. (2021). Nonverbal overload: A theoretical argument for the causes of Zoom fatigue. Technology, Mind, and Behavior, 2(1).
Zhu, R., & Argo, J.J. (2013). Exploring the impact of various shaped seating arrangements on persuasion. Journal of Consumer Research, 40(2), 336-349.
Danziger, S., Levav, J., & Avnaim-Pesso, L. (2011). Extraneous factors in judicial decisions. PNAS, 108(17), 6889-6892; and the critiques: Weinshall-Margel, K., & Shapard, J. (2011). Overlooked factors in the analysis of parole decisions. PNAS, 108(42), E833; and Glöckner, A. (2016). The irrational hungry judge effect revisited: Simulations reveal that the magnitude of the effect is overestimated. Judgment and Decision Making, 11(6), 601-610.
Lynott, D., Corker, K.S., Wortman, J., Connell, L., Donnellan, M.B., Lucas, R.E., & O’Brien, K. (2014). Replication of "Experiencing physical warmth promotes interpersonal warmth" by Williams and Bargh (2008). Social Psychology, 45(3), 216-222.
You can download the full PDF-version of the article here.
