A mediation rarely fails because the parties want incompatible things. It fails because one party, or both, arrives with a picture of what they can achieve that does not match what is actually available. The community believes a court will hand it a number it has never been awarded in that jurisdiction. The company believes a community can reach internal consensus on the same calendar its own legal team runs on. Neither belief is irrational. Both are sincere. And both will quietly wreck a negotiation if no one tests them.
Reality testing is the craft of helping each party examine its own assumptions before those assumptions harden into deadlock. You do not tell a party it is wrong. You ask questions that let the party check its position against evidence, against likely alternatives, and against the practical cost of walking away. Done well, the adjustment comes from the party’s own reasoning. Done badly, it reads as the mediator taking sides. This article is about doing it well in the specific conditions of mining and community disputes.
Why community mediation needs it more than most
Community disputes concentrate every condition that distorts expectations. Power is unequal. Information is unevenly held. Communication styles differ across cultures, and the timelines that govern a community’s decisions look nothing like a project financing schedule. Parties anchor on optimistic numbers not because they are foolish, but because they hold incomplete information and carry real pressure from the people they answer to.
Consider a scenario drawn from patterns across community compensation talks. A community facing the loss of grazing land opens with a demand far above any resettlement award the jurisdiction has ever produced. The leaders are not bluffing. They have heard rumors of larger awards elsewhere, they are uncertain what is genuinely achievable, and they are terrified of returning to residents having accepted too little. If the mediator simply calls the figure unrealistic, the community hears an ally of the company and digs in. The demand becomes a matter of principle rather than a number to be examined.
The opposite distortion shows up on the corporate side. Imagine a company that assumes a benefit agreement its lawyers drafted in six weeks can be ratified by a community in six weeks. The community needs six months to build genuine consensus. The mediator’s job is not to enforce either calendar. It is to help each side see the real constraints the other is living inside. As I argue in the work of separating the factual from the felt in mining disputes, most expectation gaps are a mix of bad information and unspoken fear. You cannot close them by argument alone.
These distortions are also self-reinforcing. A community that has been told the company is rich tends to read every counter-offer as a lowball, which pushes its anchor higher. A company that has met community deliberation before tends to assume delay is a tactic rather than a genuine process, which makes it push harder on the calendar. Each side interprets the other’s behavior through its own distorted lens. Reality testing interrupts that loop by replacing assumption with examined fact, one question at a time.
Five techniques that adjust expectations without alienating anyone
Reality testing works through sequenced questions, not statements. The five techniques below each open a different door, and a skilled mediator moves between them as the conversation requires.
The harm analysis question invites a party to articulate what happens if no deal is reached. You ask the community what follows if there is no agreement by the time the board meets. Does the company proceed without one, or pause? What happens to the jobs, to the revenue-sharing, to the timeline everyone was counting on? The questions are genuine. You do not know the answers until the party works them out aloud. The realization that impasse helps no one arrives from the party, not from you. That ownership is the whole point, because a conclusion a party reaches itself is far harder to retract than one a mediator hands over.
The comparable outcomes question brings evidence into the room without you arguing a case. You ask a company proposing a ten-year cash commitment how many comparable agreements in the region actually ran the full ten years without modification. You ask what communities say now about the changes that did occur. The answer usually exposes a gap the party had not examined.
The implementation feasibility question asks a party to describe, in operational detail, how its stated objective would actually function. A community that wants annual review meetings is asked who attends, how often, what counts as non-compliance, and who verifies it. Many demands soften the moment a party tries to build the machinery behind them. A community that confidently asks for independent monitoring may not have priced what independent monitoring costs or requires. The question does not weaken the demand. It makes the demand specific enough to be funded, staffed, and enforced.
The objective standards question moves parties off positional bargaining. A community asks for a three percent royalty. The company counters with one percent. Rather than splitting that gap, you ask what external benchmark should decide what is fair. Regional comparables, published government guidance, and recognized industry practice are all candidates. You set the standard first, then test both positions against it.
The alternatives reality question is where the craft connects to negotiation theory. Each party has a best alternative to a negotiated agreement, the concept Roger Fisher and William Ury introduced in Getting to Yes (1981). You ask the party to describe, concretely, what it will do if there is no deal. How long would that path take, what would it cost, and how likely is it to deliver the result the party is holding out for? The community that insists on a fixed jobs number is invited to weigh that against the real shape of protest, litigation, or appeal. That means counting the months or years those routes consume, not just imagining the verdict at the end.
A scenario: testing a local-hire demand
Imagine a mid-sized operator in mediation with communities over an expansion next to an existing mine. The communities had seen inconsistent benefit delivery the first time, so they open hard: half of the new workforce hired locally, with priority for those displaced by the original project. The company’s practice is closer to fifteen percent, since many roles need technical qualifications that are scarce locally. The company calls fifty percent infeasible. The community reads that refusal as proof the company does not care.
The mediator uses the implementation feasibility question. What would a fifty percent commitment actually look like? Who gets hired when the community lacks people with the specific skills the mine needs? Who trains them, over what period, and who pays? What happens to a worker placed in a role they cannot yet perform? As the community works through these, its expectation shifts on its own reasoning. A rigid target could put residents into jobs they would lose, and training programs take time to stand up. Year one cannot look like year three.
The community moves to a tiered position: thirty percent local in year one, forty in years two and three, with a company-funded training pipeline for high-potential members. The shift came from the community’s own reasoning, not from the mediator declaring fifty percent impossible. That distinction is what kept the community at the table rather than entrenched.
The mediator then turns the comparable outcomes question on the company. What hiring ratios did similar operations in the region actually reach? The company discovers several comparable mines achieved thirty-five to forty percent in their second and third years. Its own opening figure was low. Its expectation moves up rather than down. The agreed provision starts at twenty-five percent, climbs to forty by year three, funds training from year one, and adds a joint committee tracking both metrics quarterly. Neither opening position survived intact. The result was specific, monitorable, and matched what each side actually needed. Notice that reality testing did not split the difference. It moved the community down and the company up, toward a shared picture of what the local labor market could actually deliver.
When reality testing backfires
The technique damages credibility when it is applied clumsily, and four failure patterns recur. Watch for them in your own practice.
First, it turns accusatory. If your tone implies the party is being foolish, you get defensiveness, not reflection. The question is “help me understand how you reached this number,” never “how could you think that is realistic.”
Second, it gets one-sided. If you test the community’s figure relentlessly but never question the company’s timeline, the community correctly concludes you are biased. Every party deserves the same rigor. This matters most where power is uneven, because the weaker party is watching to see whether you reinforce the imbalance or hold both sides to account. That even-handedness is one of the core qualities that separate trained mediators from improvisers.
Third, it comes too early. If you start testing before a party has fully voiced its position and felt heard, the questions land as interruption. Let people finish. Acknowledge the concern. Then test.
Fourth, it runs on guesswork. If you ask a party to weigh its alternatives against comparable outcomes, you must be ready to supply accurate figures on what those outcomes are. Speculative reality testing is worse than none, because a single wrong number forfeits the trust the whole technique depends on. Before you raise comparables, do the research.
Load the questions before you sit down, not during the session
The Reality Testing Question Bank and Readiness Checklist, a companion tool, holds the five question sequences above in ready-to-use form, roughly 17 items across six sections. The opening section, “Readiness: do the homework first,” lists the comparables, awards, and timelines you must have on hand before you challenge a single figure. Five technique sections then give you the exact question order for harm analysis, comparable outcomes, implementation feasibility, objective standards, and alternatives. A closing “Guardrails” section catches the failure patterns that turn reality testing into perceived bias. You score each item In place, Partial, or Absent, and the rule is firm: do not open comparable-outcome questioning until the Readiness items are In place. Keep it beside you when a case has both sides anchored on numbers that cannot both be true. Download the Reality Testing Question Bank and Readiness Checklist.
Why this belongs inside a mediated, structured process
Reality testing is not a trick a party can run on the other side. It is not something a company’s own community-relations team can credibly perform on the people across the table. The moment one side starts questioning the other’s numbers, the questions read as tactics. The power of the technique depends entirely on its source being neutral. Only an independent third party can hold that neutrality, because it has no stake in the outcome and applies the same hand to both sides. Such a mediator can ask a community to examine its alternatives and ask a company to examine its assumptions in the same session. Neither side reads manipulation into it. That is the practical case for bringing in a mediator rather than negotiating expectations across an adversarial table. It is also worth weighing honestly against the question of when an in-house team can carry the conversation alone.
Reality testing also needs a frame larger than a single hard conversation. Expectations re-inflate the moment people return to their constituents. The adjustments have to be captured, documented, and carried into an agreement that holds. That is the role of the Social Accord Architecture, the methodology I use to turn tested, realistic expectations into durable commitments. Within the Social Accord Architecture, reality testing is not a one-time intervention but a continuous discipline. It runs through how positions are surfaced, how alternatives are examined, and how the resulting accord is structured to survive implementation. The SAA gives the technique somewhere to land.
The one thing to take into your next session
Before you challenge a single number, do the homework the numbers require. The most common reason reality testing fails is not poor questioning. It is a mediator who invites a party to weigh its alternatives, then cannot say what comparable awards, hiring ratios, or timelines actually look like in that context. Build the evidence base first. Then ask the questions that let each party adjust on its own reasoning. Apply the same rigor to the powerful side and the less powerful side. Capture what shifts before it drifts back. If you want a second view on a case where expectations are stuck on both sides, write to me at thomas@thomasgaultier.com.
