# Evidence and sources

Research checked 10 September 2026. This is a focused research synthesis, not a systematic review. Searches used Firecrawl’s paper index, citation-related expansion and online primary sources. Full text was not available for every paper. These original workflows combine research-informed ideas with practical design; no study evaluated this pack, its AI prompts, or its effectiveness in agencies of 20–200 people.

## What the evidence changes

Do not increase conflict as a performance tactic. Meta-analyses provide substantial counterevidence to that advice, including the 2026 update. Invite relevant information and real dissent while limiting personal conflict, unnecessary escalation and procedural confusion.

Understanding another perspective requires checking with the person. Imagining their motives can increase confidence without improving accuracy. Listening may improve the quality of contact without changing beliefs. A fair decision can leave people in disagreement.

Power is part of the process. A junior person’s apparent agreement may reflect consequences of dissent. Safe participation routes, an independent route for serious allegations and explicit decision rights are design requirements, not guarantees of psychological safety.

## Source ledger

The skill references below use shortened endings of `disagreement-…` names.

| ID and academic source | Evidence and access | Finding used, limit and application |
|---|---|---|
| D1. De Dreu & Weingart (2003), *Task versus relationship conflict, team performance, and team member satisfaction: A meta-analysis*. [DOI](https://doi.org/10.1037/0021-9010.88.4.741); [primary paper copy](https://web.mit.edu/curhan/www/docs/Articles/15341_Readings/Negotiation_and_Conflict_Management/De_Dreu_Weingart_Task-conflict_Meta-analysis.pdf). | Meta-analysis; full text accessible and discussion checked. | Both conflict types showed negative associations with performance. Associations do not establish that removing every disagreement causes improvement. Informs issue-map, surface-dissent and reset-exchange: gather information without encouraging conflict for its own sake. |
| D2. de Wit, Greer & Jehn (2012), *The paradox of intragroup conflict: A meta-analysis*. [Institutional record](https://repub.eur.nl/pub/37902); [DOI](https://doi.org/10.1037/a0024844). | Meta-analysis of 116 studies, 8,880 groups; abstract/institutional summary checked. Firecrawl body unavailable. | Context matters; task conflict results depend on conditions and outcome definition. Moderation is not a recipe guaranteeing benefit. Used with D1 and D3 in issue-map, surface-dissent and reset-exchange. |
| D3. Yuan, Yin & Sun (2026), *The paradox of team conflict revisited: An updated meta-analysis of the team conflict–team performance relationships*. [DOI](https://doi.org/10.1037/apl0001315); [PubMed](https://pubmed.ncbi.nlm.nih.gov/40875336/). | Updated psychometric meta-analysis; abstract and metadata verified. Online 2025, journal issue 2026. No full text retrieved. | Task, relationship, process and status conflict are negatively related to performance on average; substantial heterogeneity remains. This is counterevidence to “healthy conflict always improves teams,” not a causal test of silence. Informs issue-map and surface-dissent. |
| D4. Itzchakov, Weinstein, Leary, Saluk & Amar (2024), *Listening to understand: The role of high-quality listening on speakers’ attitude depolarization during disagreements*. [DOI](https://doi.org/10.1037/pspa0000366); [paper](https://perc.wa.gov/wp-content/uploads/2024/07/Listening_Depolirization.pdf). | Experimental series; full-text results and moderation limits checked. | Attentive, understanding listening supported less defensive reflection and some depolarization outcomes. Effects varied by measure; this does not mean accepting a particular workplace proposal. Informs listen-and-check and reset-exchange. |
| D5. Santoro, Broockman, Kalla & Porat (2025), *Listen for a change? A longitudinal field experiment on listening’s potential to enhance persuasion*. [DOI/full paper](https://doi.org/10.1073/pnas.2421982122). | Preregistered field experiment, 1,485 completed post-conversation measures; publisher discussion/results checked. | Adding listening to a narrative appeal did not improve persuasion, though interpersonal perceptions improved. One contentious policy, video conversations and willing participants limit transfer. Informs listen-and-check: understanding and cooperation, not guaranteed belief change. |
| D6. Yeomans, Minson, Collins, Chen & Gino (2020), *Conversational receptiveness: Improving engagement with opposing views*. [DOI](https://doi.org/10.1016/j.obhdp.2020.03.011); [author copy](https://www.mikeyeomans.info/papers/receptiveness.pdf). | Experiments and observational field evidence; publisher abstract and study sections checked. | Language expressing willingness to engage affects perceived receptiveness; field escalation associations are not randomized effects. Clear-dissent and reset-exchange use open questions and calibrated claims without treating politeness as proof of fairness. |
| D7. Eyal, Steffel & Epley (2018), *Perspective mistaking: Accurately understanding the mind of another requires getting perspective, not taking perspective*. [DOI](https://doi.org/10.1037/pspa0000115); [abstract](https://pubmed.ncbi.nlm.nih.gov/29620401/). | 25 experiments; abstract checked. | Instructed perspective taking did not consistently improve accuracy; direct conversation improved accuracy in a final experiment. Different tasks limit blanket conclusions. Listen-and-check and fact-check distinguish speculation from information obtained. |
| D8. Schulz-Hardt, Jochims & Frey (2002), *Productive conflict in group decision making: Genuine and contrived dissent as strategies to counteract biased information seeking*. [DOI/publisher](https://doi.org/10.1016/S0749-5978(02)00001-8). | Experiment with 201 employees/managers; publisher abstract checked. | Genuine preference heterogeneity was more effective against confirmatory information search than assigned devil’s advocacy. This does not prove all dissent is useful. Surface-dissent distinguishes real objections from AI-generated or assigned challenges. |
| D9. Colquitt, Conlon, Wesson, Porter & Ng (2001), *Justice at the millennium: A meta-analytic review of 25 years of organizational justice research*. [DOI](https://doi.org/10.1037/0021-9010.86.3.425); [university paper](https://leeds-faculty.colorado.edu/dahe7472/Colquitt%202001.pdf). | Meta-analysis of 183 studies; abstract/intro checked. | Distinct fairness dimensions contribute to fairness perceptions. Much underlying research is associational. Fair-process, close-decision and review-resolution separate fair treatment, process and outcomes. No fairness score or legal determination is inferred. |
| D10. Behfar, Peterson, Mannix & Trochim (2008), *The critical role of conflict resolution in teams: A close look at the links between conflict type, conflict management strategies, and team outcomes*. [DOI](https://doi.org/10.1037/0021-9010.93.1.170); [author paper](https://www.billtrochim.net/research/JAP%20Conflict%20Resolution%202008.pdf). | Qualitative study of 57 autonomous teams; abstract checked. A published correction was surfaced; no corrected table estimates are used here. | Conflict-management approaches and explicit reasons for work decisions distinguished team patterns. Observational patterns are not proof of intervention effects. Fair-process, close-decision and review-resolution use explicit reasons and follow-through. |
| D11. de Dreu, Koole & Steinel (2000), *Unfixing the fixed pie: A motivated information-processing approach to integrative negotiation*. [DOI](https://doi.org/10.1037/0022-3514.79.6.975); [university paper](https://research.vu.nl/ws/files/1681682/dreu%20Journal%20of%20Personality%20and%20Social%20Psychology%2079%286%29%202000%20u.pdf). | Two experiments; abstract and research design checked. | Process accountability reduced fixed-pie perceptions and supported integrative outcomes under studied conditions. Simulated negotiation does not guarantee joint gains at work. Fact-check and integrative-options test priorities and actual trade-offs. |
| D12. Galinsky, Magee, Inesi & Gruenfeld (2006), *Power and perspectives not taken*. [DOI](https://doi.org/10.1111/j.1467-9280.2006.01824.x); [university copy](https://web.mit.edu/curhan/www/docs/Articles/15341_Readings/Power/Galinsky_et_al_Power-and-Perspectives-Not-Taken.pdf). | Four experiments and a correlational study; abstract checked. | Power was linked with reduced spontaneous perspective taking in these tasks. It is not a diagnostic claim about every manager. Clear-dissent and fair-process explicitly inspect access, authority and consequences rather than presume equal participation. |
| D13. Kong, Dirks & Ferrin (2014), *Interpersonal trust within negotiations: Meta-analytic evidence, critical contingencies, and directions for future research*. [DOI/publisher](https://doi.org/10.5465/amj.2012.0461). | Meta-analysis, 38 independent studies; abstract checked. | Trust’s relationship with negotiation processes/outcomes is contingent. The pack does not prescribe unconditional disclosure or assume trust is sufficient. Integrative-options and close-decision make constraints and commitments explicit. |
| D14. Galinsky, Maddux, Gilin & White (2008), *Why it pays to get inside the head of your opponent: The differential effects of perspective taking and empathy in negotiations*. [Primary paper](https://willmaddux.web.unc.edu/wp-content/uploads/sites/15846/2019/01/Psych-Science-PT-Negotiations.pdf); [DOI](https://doi.org/10.1111/j.1467-9280.2008.02096.x). | Negotiation experiments; abstract and author copy opening checked. | Perspective taking can aid negotiation under studied conditions. Read alongside D7: imagining priorities is not accurate knowledge. Integrative-options generates hypotheses and asks the counterpart to confirm priorities. |

## Original design choices and limits

The issue categories, worksheets, output schemas, timing suggestions, stop conditions and AI interaction pattern are original practice design. They have not been evaluated as a package. Serious harm allegations are outside ordinary disagreement facilitation; the independent-route safeguards are prudent boundaries, not an empirically validated reporting protocol.

Many studies use students, simulations, self-report or settings outside small professional-services businesses. Meta-analyses aggregate heterogeneous measures and do not automatically eliminate confounding, reverse causation or publication bias. The source set was selected for relevance and opposing findings, not exhaustiveness. No study or author endorses this pack.