Difficult Conversations: Summary, and What the Research Says About Hard Talks at Work

What the book argues

Difficult Conversations: How to Discuss What Matters Most is by Douglas Stone, Bruce Patton and Sheila Heen. Viking published it in 1999 (Publishers Weekly, 1999; Open Library, n.d.-a). A 2010 Penguin edition added a new chapter (Open Library, n.d.-b). Penguin Books published a third edition, with a foreword by Roger Fisher, on August 22, 2023, which the publisher says addresses race, culture, gender, power, social media and communicating through technology (Penguin Random House, n.d.).

Publishers Weekly described the authors as teaching at Harvard Law School and the Harvard Negotiation Project, and the book as drawing on law, organizational behavior, psychology and dialogue studies (Publishers Weekly, 1999). Its core idea, as a summary by the Conflict Research Consortium explains, is that underneath every difficult conversation are three others (Beyond Intractability, n.d.). The "What happened?" conversation is about disagreement over the facts, what should happen and who is to blame. The feelings conversation is about each person's emotions and whether they are valid. The identity conversation is the internal one, about what the situation says about who you are. The summary reports the authors' warning that focusing on blame is a bad idea because it inhibits learning what is really causing the problem; the alternative is to map each person's contribution (Beyond Intractability, n.d.).

In my words: most hard talks go wrong because each person is defending a version of events, a feeling they have not named and a picture of themselves, all at once.

Who it is for

Managers who avoid the conversations a change requires: telling someone their role is changing, raising a problem with their own boss, or admitting a decision went badly. Publishers Weekly called it a cogent resource for those who see the sense in preparing for tough talks in advance, while noting that many of its points are simplistic, though often overlooked in stressful moments (Publishers Weekly, 1999).

What the research says

The book is a practical guide built from teaching and practice. I found no trial of its method in the evidence I use. Its cautions are close to research on speaking up and on feedback.

Raising concerns is linked to lower job performance. Chamberlin, Newton and LePine (2017) pooled research on employee voice. According to their abstract, promotive voice, suggesting improvements, was positively associated with job performance, while prohibitive voice, raising concerns about problems, was negatively associated with it. That is correlational, but it fits the book's premise that the conversation about what is going wrong is the hard one.

Stress goes with less speaking up. Ng and Feldman (2012), according to their abstract, found that workplace stress was negatively related to voice, and that voice was positively related to performance outcomes. That is a possible reason to have the hard conversation before a change reaches peak pressure, though the studies cannot show that stress comes first.

Feedback can backfire when attention shifts to the self. Kluger and DeNisi (1996) pooled studies of feedback interventions. According to their abstract, feedback improved performance on average (d = .41), but more than a third of the interventions lowered performance, and feedback became less effective as attention moved away from the task and toward the self. That loosely fits the book's identity conversation: a message that draws attention to the self may pull attention away from the task.

Emotional skill and constructive conflict. Schlaerth, Ensari and Christian (2013) pooled 20 studies. According to their abstract, emotional intelligence was positively associated with constructive conflict management, more strongly for subordinates than for leaders. It supports the book's attention to the feelings conversation, as association.

Where the evidence is thin

No study I use tests the three-conversations framework or the step-by-step method. The research above is mostly correlational and measures voice, feedback and conflict styles, not conversations handled the book's way. I also found no evidence on the book's specific advice to map contribution instead of blame.

How it maps to the Unfrozen Loop

This mapping is my judgment. A difficult conversation is how Surface both ways happens in practice: the part of the Unfrozen Loop where bad news travels up as easily as plans travel down depends on someone starting the talk. The conversations between a manager and the people above, about what a change will cost, belong to Make sense together, top included. Making those talks safe is the culture ring: safe to speak, and speaking changes something. See the Atlas answer on why bad news does not reach leaders.

What I take from it

The book's three conversations are a useful way to prepare. Before a hard talk about a change, write one line for each: what you think happened, what you feel about it, and what you are afraid it says about you. Then ask the other person for theirs. For a related book on high-stakes talks, see Crucial Conversations, and for direct feedback, Radical Candor.