Defending Objectivity and Conducting Trustworthy and Trusted Analyses

6. Civic Knowledge and Skills

This introductory article was written by ChatGPT at the direction of Heidi Burgess, who reviewed, edited, and approved the final content.
For more information

June 24, 2026

Many of today’s most difficult social and political conflicts depend heavily on technical facts that ordinary citizens cannot easily determine for themselves. Climate change, public health, crime policy, education, immigration, election administration, environmental regulation, economic inequality, and many other issues all require specialized knowledge. Citizens can observe parts of these problems directly, but those observations are almost always partial. People need experts to gather data, test explanations, compare evidence, and explain what is known, what is uncertain, and what remains genuinely disputed. Without trustworthy analysis, public debate is easily captured by rumor, ideology, anecdotes, and politically convenient claims.

This is why defending objectivity is so important. Objectivity does not mean that researchers have no values, no assumptions, or no personal views. It means that they use disciplined methods designed to keep those views from determining the results. The National Academies’ report on research integrity describes the core values of good research as objectivity, honesty, openness, fairness, accountability, and stewardship. These values are not abstract virtues. They are practical safeguards against self-deception and manipulation. Researchers need clear methods, reliable measurements, honest statistics, disclosure of conflicts of interest, peer review, replication where possible, and a willingness to correct errors. These practices do not make science perfect, but they make it much more reliable than unsupported assertion or individual intuition.

The credibility of expertise has been damaged, however, by real failures. Some researchers have engaged in misconduct, such as fabrication, falsification, or plagiarism. Others have used weaker, but still harmful, practices, such as selective reporting, hiding inconvenient data, overstating findings, or framing results in ways that support a political, financial, or professional interests. Even when misconduct is rare, highly publicized cases can tarnish the reputation of entire fields. The problem is made worse when experts speak with more certainty than the evidence justifies, when preliminary findings are presented as settled conclusions, or when scientific institutions appear to punish dissent, rather than answer it. In polarized settings, these failures allow people to dismiss not just a particular study, but “science” or “expertise” as a whole.

The answer is not to abandon expertise, but to make expert analysis actually and visibly. The National Academies’ work on reproducibility and replicability emphasizes the importance of methods that allow others to check results, repeat analyses, and test whether findings hold up under different conditions. In many fields, this means pre-registering studies, sharing data and code when privacy and security allow, clearly separating exploratory from confirmatory findings, and publishing negative or mixed results, rather than only exciting ones. It also means rewarding careful work, not just dramatic claims. Good analysis should make it easier for critics to see exactly how conclusions were reached and where reasonable disagreement remains.

Experts also have to communicate in ways that earn public trust. This does not mean oversimplifying complex issues or pretending that evidence can answer every moral and political question. It means explaining findings in plain language, acknowledging uncertainty, and being clear about the difference between scientific conclusions and policy recommendations. The National Academies’ report on communicating science effectively stresses that science communication should be grounded in an understanding of audience concerns, values, and prior beliefs. People are more likely to listen when experts answer the questions regular citizens are actually asking, admit the limits of their knowledge, and avoid treating skepticism as stupidity. Trust grows when experts are both competent and respectful.

In conflicts where facts themselves are disputed, the process for producing analysis matters almost as much as the analysis itself. Our materials on joint fact-finding and obtaining trustworthy information show how contending parties can sometimes work together to choose experts, define questions, review evidence, and identify points of agreement and uncertainty. This does not guarantee consensus, and it is not possible in every conflict. But it can reduce the use of dueling experts who simply reinforce the positions of their sponsors. When people have had a fair role in shaping the inquiry, they are more likely to accept its results, even if those results are uncomfortable.

Defending objectivity is therefore a civic responsibility, not just a professional one. Experts should protect the integrity of their methods and resist pressure to become advocates disguised as analysts. Journalists should report scientific findings with context, not just dramatic headlines. Policymakers should seek independent expertise and avoid attacking analysts simply because their findings are inconvenient. Citizens should be skeptical in the best sense of the word: asking for evidence, checking sources, and distinguishing serious expert disagreement from politically motivated denial. Complex societies cannot solve complex problems if they reject expertise altogether. They need analysis that is rigorous enough to be trustworthy and transparent enough to be trusted.

—————————

This page was created by ChatGPT in response to this prompt. It was then reviewed, edited, supplemented and approved by Heidi Burgess. More information about how and why we are using AI in this way, and about the growing number of ways in which Beyond Intractability is using ChatGPT, Claude and other AI systems to generate content and build out the BI system, is available on our BI/AI Overview Page.