Can a Tabletop RPG Really Teach Conflict Skills? What the Research Says

There's a scene in Operation Aetherfall – the training scenario we've now run through two pilot sessions – where the players bring their commanding officer intelligence that contradicts his plan. Commander Korr doesn't yell. He doesn't pull rank. He says, "Until I see formal revision, I'm proceeding on the validated model." And the players have to decide, right there at the table, whether to keep pushing a superior who sounds entirely reasonable and is entirely wrong.

Nobody at your company talks like a fantasy military officer. Every leader you've ever worked for has said some version of that sentence.

So: can a tabletop RPG really teach workplace conflict skills? The research says yes, with conditions. Practice-based training outperforms lectures for interpersonal skills, and the fantasy frame has a documented psychological mechanism behind it – it's not decoration. But the conditions have teeth. The skills have to live inside the game's win condition. The debrief carries most of the learning. And one session, no matter how good, won't rewire anyone. Here's the evidence as we read it, including the parts that don't flatter the format.

Does practice actually beat a lecture?

For interpersonal skills, yes – and the pattern holds across every field that's studied it. Military after-action reviews, medical crisis simulation, business negotiation exercises: the research from all three converges on the same finding. People build interpersonal skills by working through authentic dilemmas under pressure, then reflecting on what happened in a structured way. A 2020 meta-analysis of 61 studies covering more than 3,000 people (Keiser & Arthur) found that after-action reviews – the structured debriefs military teams run after an exercise – improve performance by d = 0.79. In training research, that's a large effect. Most interventions never get close.

Conflict skills specifically sort along a gradient of teachability. Interest-based negotiation is the most trainable: a meta-analysis of 17 studies found a large effect (Hedges' g = .72), and more than 80% of trained professionals reported still using the skills in their work long afterward. Consensus-building responds well too, at least at the procedural level – you can teach a group to follow a decision process in a day. Managing up and overcoming conflict avoidance sit at the hard end. Both involve emotional regulation and threat response, and those don't shift in a single session.

That gradient matters when you're deciding what to promise a team. A well-designed session can teach people to ask interest-based questions and run a real decision process. It can only begin the work of making anyone comfortable in conflict.

How can a fantasy mission train real meeting behavior?

Because skill transfer follows psychological fidelity, not surface realism. Psychological fidelity is the degree to which a simulation evokes the same cognitive and emotional processes as the real situation – the same pressure, the same incomplete information, the same discomfort about challenging someone with more power. Decades of simulation research across military, medical, and business settings find that this is what drives transfer. Physical resemblance to the workplace barely registers. A 2024 study (Davies & Krame) found novices actually learn better in lower-fidelity environments that isolate the key cognitive demands, and a 2024 study on abstraction in serious games concluded that fantasy framing is "not a compromise in fidelity but a deliberate strategy to support transfer."

The transfer-of-learning literature backs this from the other side: structural similarity between training and application matters more than surface similarity. When a player presses Korr about his intelligence sources, the surface is pure fantasy. The structure is a meeting you've sat in. At work, leaders rarely say "because I said so." They say "because the data says so," and you have to challenge the data without torching the relationship. That's the exact task the scenario demands, in costume.

The skin of the scenario is arbitrary. The pressure isn't.

Won't adults find this silly?

Some will, at first – and the mild silliness is doing real work. Drama therapy has a name for the mechanism: aesthetic distance, introduced by Robert Landy in 1983. When you act through a character, you can attribute the uncomfortable move to the character instead of yourself. That wasn't me arguing with the commander. That was Bramm. The distance lowers the identity threat that normally stops adults from experimenting with behavior they're bad at, which is why the awkwardness of playing a character is a feature and not an obstacle. Augusto Boal built his Forum Theatre on the same insight in the 1970s and said it plainly: "The theatre itself is not revolutionary; but it offers a chance to rehearse for revolution."

The research also names the failure modes, and they're symmetric. Underdistance: the material cuts too close, and participants shut down or get hurt. Overdistance: the fiction drifts so far from reality that nothing transfers – people had fun and learned nothing. Role-play researcher Sarah Lynne Bowman calls the spillover between what the character experiences and what the player feels "bleed," and bleed is the mechanism by which practice at the table becomes change at work. A session with zero discomfort probably isn't producing any.

Two things the frame can't survive. First, an unsafe group. Amy Edmondson's research defines psychological safety as "a shared belief that the team is safe for interpersonal risk taking," and it's a group-level property – if people fear real judgment from real colleagues, the protective fiction collapses. Second, coercion. Mollick and Rothbard found that consented gamification increases positive affect while mandated gamification decreases it and can reduce performance. Dobbin and Kalev's analysis of 829 firms found mandatory diversity training decreased representation of Black women in management by 9%. Voluntary participation is a design requirement, not a courtesy.

What keeps this from being a lecture with dice?

Intrinsic integration: the skill being practiced has to be how you win, not a worksheet stapled to the fun. Serious-games researchers Habgood and Ainsworth found that learning embedded in core mechanics works through attention – "players only pay attention to features needed for the game task." If the conflict skill is the game task, people practice it because the mission demands it. If it's bolted on top – points, badges, a leaderboard over a lecture – it backfires, draining intrinsic motivation through what psychologists call the overjustification effect.

Here's what integration looks like in practice. In Operation Aetherfall, each of the six players holds a piece of secret information and a private motivation. The team literally cannot complete the mission without surfacing what everyone knows, which makes interest-based questioning the winning move rather than the assigned exercise. Hidden information also blocks what game designers call "quarterbacking" – one confident player directing everyone else – because nobody holds enough of the picture to steer alone. And the motivations pull against each other by design: one player's card rewards urgency, another's rewards humanitarian conscience. The disagreement that erupts over the mission timeline isn't scripted. It's structural.

Building this mapping is the genuinely hard part of the format. The Learning Mechanics–Game Mechanics framework (Arnab et al., 2015) exists because translating a learning objective into a game mechanic is design work, not garnish. We've written up the practical side in how to run a tabletop RPG training session.

Where is the evidence thin?

Thinner than the enthusiasm, and it's worth being precise about where. Two scoping reviews anchor the TTRPG-specific evidence. Arenas, Viduani, and Araujo (2022) screened a pool of 4,069 studies down to 50 sources and found consistent positive outcomes across psychological well-being, social anxiety, and resilience. Yeo and Tan (2024) reviewed ten years of research and reached similar conclusions for social skills, perspective-taking, and emotional regulation. Consistent and promising. Also: only 16% of the studies in the 2022 review were experimental. Rigorous randomized trials remain rare.

The bigger caveat is that the evidence base is largely analogical. The strong effect sizes come from adjacent fields – drama therapy, medical simulation, military after-action reviews – rather than from tabletop RPGs run as corporate training. No published study has tested TTRPG-based conflict training for workplace teams specifically. When we summarized our own research corpus, we wrote that the approach is "evidence-informed but not evidence-replicated," and that's the sentence anyone evaluating this format deserves to see first.

Engagement doesn't equal learning, either. A JMIR meta-analysis found gamified tasks were more motivating but no better for actual performance – you can fill a room with laughter and send everyone home unchanged. And for some people this is simply the wrong format: participants with social anxiety or trauma histories can be harmed by role-play they didn't choose. We wrote when game-based training is the wrong tool to map those boundaries, and we mean it as a checklist, not a disclaimer.

What can one session actually change?

Less than most training marketing implies. By default, only 10-15% of trained skills transfer to workplace performance at all. Interpersonal skills decay within 30-90 days without practice. In the CPP Global study, only 27% of trainees reported feeling more comfortable with conflict after training. One study of 234 participants found conflict training actually reduced willingness to confront among high-avoidance individuals – the people who most needed help moved backward.

What moves those numbers sits after the game ends. The debrief carries the load: every validated framework follows the same three-phase arc of reactions, analysis, and application, and we walk through it in how to debrief a training exercise. Then come implementation intentions – structured if-then plans formed before anyone leaves. A 2025 meta-analysis of 642 tests (Sheeran, Listrom, & Gollwitzer) found behavioral effects of d = .66 when plans use the if-then format, and in one study participants who formed them transferred their training at 67% against the standard 30%. Written commitments shared with an accountability partner produced 76% goal achievement versus 43% for unshared goals (Matthews, 2015). In Hughes et al.'s 2020 meta-analysis, peer support predicted sustained transfer more strongly than supervisor or organizational support. Skills stick where colleagues expect to see them used. The rest of that story is in why workplace training doesn't stick.

Our own first pilot got this wrong. The debrief section in the original adventure document was three questions. Three – set against a research base saying the debrief is where the learning actually happens. So we rebuilt the session around that finding. The second pilot's run-of-show gives a structured 90-minute debrief inside a seven-and-a-half-hour day, three times the slot we give to teaching the rules. Between pilots we also cut the adventure from five acts to three and the dungeon from six rooms to four, because the training-relevant moments were drowning in dungeon.

So what's the honest verdict?

I run a consulting firm, not a game store. If the dice didn't hold up against the transfer literature, we'd be selling something else. What the evidence supports is narrower and better than the hype: a well-built scenario generates authentic conflict to practice on, the fantasy frame lowers the cost of experimenting, the debrief converts the experience into something usable, and the weeks afterward decide whether any of it sticks. Buy the format for what it does. Plan for what it doesn't.

Commander Korr isn't real. The skill of pushing back on him is.