News Is More Than Facts: The Moral Foundations Behind Media Framing

When readers compare coverage of the same news event across different outlets, they often notice that, even when the core facts are similar, the meaning feels different. 

One article emphasizes suffering and compassion. Another stresses law, order, and stability. A third focuses on justice and equality. Readers sometimes interpret these differences as distortion or bias. But often, the difference is not the presented facts, but in the moral emphasis. 

To understand this dynamic, we need a framework that goes deeper than left vs. right labels. The Moral Foundations Theory offers such a lens.

The Five Moral Foundations

The Moral Foundations Theory, developed in moral psychology and popularized by the book The Righteous Mind, proposes that human moral reasoning is built on several core foundations. The original framework identifies five:

• Care / Harm – concerns about suffering, compassion, and protection of the vulnerable
• Fairness / Cheating – concerns about justice, equality, and reciprocity
• Loyalty / Betrayal – concerns about group belonging and solidarity
• Authority / Subversion – concerns about respect for institutions and social order
• Sanctity / Degradation – concerns about purity, dignity, and bodily integrity

Extensive experiments have shown that these foundations are not distributed evenly across the political spectrum. In the United States and other Western democracies, people who identify as politically liberal tend to prioritize Care and Fairness more strongly, while people who identify as politically conservative tend to draw more evenly across all five foundations. This does not mean that one side lacks certain moral concerns. Rather, it suggests that political disagreement often reflects differences in moral weighting. 

Consider immigration debates. Some coverage foregrounds Care by focusing on vulnerable families and humanitarian obligations. Other coverage emphasizes Loyalty, highlighting national identity and obligations to fellow citizens. Other reporting stresses Fairness, asking who bears economic costs and who benefits. Political conflict, thus, is not a clash between morality and immorality. It is a disagreement about which moral foundations deserve priority in a given situation.

How Moral Foundations Appear in News

News articles rarely state their moral assumptions directly. Instead, moral framing operates through subtle cues. Word choices signal who is a victim and who is responsible. Certain harms are described in vivid detail while others receive brief mention. Particular institutions are portrayed as protectors or as threats. Some social groups are centered in the story, while others remain peripheral.

Two outlets may report accurately on the same protest. One might concentrate on property damage and the strain on law enforcement, activating Authority and Loyalty concerns. Another might focus on claims of injustice and community grievances, activating Fairness and Care. Readers experience these as different stories, even when the underlying events overlap.

Understanding moral foundations in news coverage helps clarify why disagreement persists even when facts are shared. People can accept the same evidence yet disagree about what matters most. A reader who prioritizes Authority may view a situation primarily in terms of order and institutional stability. A reader who prioritizes Fairness may focus on rights and equal treatment. Both may reason consistently within their moral framework.

Making these foundations visible encourages reflection rather than dismissal. Instead of assuming that an opposing outlet is simply misleading, readers can ask which moral values are being emphasized and which are less visible. This promotes a more self-aware form of media consumption and opens new possibilities for deeper understanding. Classifications such as left, center, and right provide useful orientation, but moral foundations offer a more nuanced insight into how different outlets construct meaning around events.

From Theory to Research

Over the past decade, researchers have increasingly studied how moral foundations appear in political and news texts. Early work developed dictionary-based approaches to detect moral language in speeches, debates, and social media, providing the first tools for mapping Care, Fairness, Loyalty, Authority, and Sanctity in public discourse. Subsequent research introduced annotated corpora of news articles, where human annotators label keywords and passages according to the five foundations. These resources have enabled the development and evaluation of machine learning models that identify moral framing more precisely and across longer narratives, including full news articles rather than short social media posts.

In my own recent work on moral frame preserving news summarization, we examine how LLMs handle moral content when generating summaries of news articles. We found that unchecked summarization approaches tend to preserve factual information while diluting or shifting moral framing. This means that summaries may appear neutral while subtly altering the value structure of the original article. Our research proposes methods to better preserve the original moral framing, highlighting that moral content is not a spurious feature but a central component of news articles.

Taken together, this line of research shows that moral foundations can be identified, measured, and compared across large collections of articles. The goal is not to rank outlets morally or to declare one foundation superior. Instead, it is to map patterns of moral emphasis in a systematic and transparent way.

Imagine reading multiple articles about the same issue and being able to see which foundations are most prominent in each. One article might lean heavily on Care and Fairness. Another might emphasize Authority and Loyalty. A third might foreground Sanctity concerns. Such information would not replace ideological labels, but it would deepen them. It would make visible the moral architecture that underlies public disagreement.

Toward More Constructive Media Consumption

News is never just a collection of facts. It is a structured narrative shaped by moral concerns that guide attention, language, and emphasis. When readers encounter coverage that feels biased, they may be reacting to a moral framework that differs from their own. If we want healthier public discourse, we need tools that make these frameworks visible. The conversation then shifts from accusations of bad faith to questions about competing values. By making those concerns more transparent, we can better understand not only our media landscape, but also one another.

Enrico Liscio is a postdoctoral researcher at TU Delft in the Netherlands. His work focuses on helping AI systems truly understand people so they can be genuinely useful in real-world settings. This challenge—known as value alignment—is about ensuring AI can recognize and respond to human values, rather than remaining at a surface-level understanding of human preferences or instructions. He has a Lean Left bias.

This blog was reviewed by Editor-in-chief Henry A. Brechter (Center).