The bedrock principle of data science is that organizations collect, process, and analyze data to make better decisions. But, according to the 2025 Data Universe Manifesto—the document guiding the direction and content attendees will experience at the second annual Data Universe in New York City in the spring of 2025—even with perfect information, organizations often struggle to turn it into meaningful action.

Why? Because, according to science journalist and author David McRaney, people don’t like to change their minds—the exact condition that data-driven thinking is supposed to encourage.

The Manifesto, written by Data Universe Conference Chair Alistair Croll, cites McRaney’s work when it addresses obstacles that prevent groups and organizations from acting on the insights drawn from good data. Hidden motivations, groupthink, fear of ostracism and competing interests all hinder logical decision-making and contribute to an inability for even rational people to change their minds.

Changing your mind, McRaney said during a keynote address at Data Universe in 2024, means acknowledging you were wrong about something. And, according to research he cited in his seminal 2022 book, How Minds Change, individual and group psychology make that very challenging.

“It's very easy for any culture to become one in which it's difficult to admit that you're wrong about something,” McRaney said during his talk. “You don't make the decision that is ‘best.’ You always make the decision that is easiest to justify. And if there is something that you would like to do or think or feel or some belief you would like to maintain, you will find a reason to do it.”

Successful, data-driven organizations find ways to combat these manifestations of human nature. They are not afraid to take risks, because mistakes are not harshly punished. Employees of successful organizations have the freedom to be wrong without shame—they have systems in place to understand where and why they are wrong and, subsequently, correct those things.

The most significant component of those systems? Developing the capacity to effectively argue.

Argue with me

It’s notoriously difficult, when there is a conflict with another person or group, to know who’s right. When someone tells you something that doesn't match your current model of reality—communicates a conclusion that you wouldn't have drawn, based on a set of data, for instance—you either must update your ideas and your trust in that person, or update your beliefs in your model of reality. In a fair system, you do that by asking the person to describe their reasoning and using the response to update your priors or not. Then you provide your response and deliberate to find a solution. The problem, McRaney says, is that we don’t evaluate internal and external reasoning fairly.

“The research is pretty strong on this,” he notes. “We are great at picking apart other people's arguments. We are terrible at picking apart our own. We evaluate our own reasoning in a biased, confirmatory manner. We evaluate the reasoning of others in an effortful, disconfirmatory manner.”

Overcoming that bias, acknowledging that they could be wrong and fairly evaluating the arguments of others are hallmarks of members of healthy teams. Such teams are more easily able to surmount the obstacles often displayed by otherwise rational humans detailed by Croll in the Data Universe Manifesto. But how do you get there?

As it turns out, once people begin sharing their conclusions with someone else—and the reasoning they used to reach them—they begin to see that they might be wrong about things. But, once you begin to let people argue in groups, that process gets supercharged. Group discussions begin to zero in on the right answers more efficiently.

“The research suggests that if you get in groups of just three or more to discuss any issue, almost everything we've ever written about on this topic flips over to, ‘humans are pretty great at this.’ You just have to let them argue,” he says. “You've got to figure out the right way to argue in your organization.”

As an example, McRaney noted an experiment that asked a room full of people a challenging trick question. When the people in the experiment were asked to come up with the answer on their own, 90 percent of the room had the incorrect answer. Then they were encouraged to talk to their neighbors. In the first 10 minutes, enough of the people who had the correct answer were able to convince 40 percent of the room to change their mind. After 20 more minutes of discussion, 100 percent of the room knew the correct answer. People can change their minds when presented with intelligent reasoning, researchers saw.

McRaney related three tips to the Data Universe audience they could use to enshrine argument as part of their corporate culture, fostering an environment in which it’s acceptable to advance ideas or strategies that might be wrong with the understanding that open dialogue will help teams reach the “right” answer.

Organize a regular “argument forum”: In these scheduled sessions, participants are separated into groups of three to five and expected to present ideas along with their reasoning and openly disagree with one another.

Designate a Devil’s Advocate: Have a person for each session who tries to pick apart the idea under examination. Make sure to rotate that responsibility to reduce the risk of any one person always being the one to call out “bad ideas.”

Praise the Devil’s Advocate: The most important thing you can do in your organization is praise the people who took on the prevailing wisdom and destroyed a potentially bad idea. You want to be wrong, in a team-oriented way, and arrive at the right answer with the same feeling.

“Don't butt heads,” he warns. “Discuss how, while it's weird that we both disagree, wouldn’t it be great to solve the mystery of why we disagree. You go shoulder-to-shoulder and look at the disagreement, entertain the idea that you're both wrong. Or, in a weird way, maybe you're both right. Your perspective and their perspective are different because the thing you're looking at can only be seen from two different sides for you to make sense of it. So, what I'm asking you to do is argue. But argue well.”