Use Scientific Methods to Detect Fake News

The principles of science can help detect false information.

Posted Nov 18, 2017

Both fake news and science became salient issues during last year’s presidential election. Although the topic of fake news has elicited research attention in several domains (e.g., political sciences, communication study, and social psychology; see De keersmaecker & Roets, 2017), it has not been examined from the perspective of scientific methods, which are one of three main components of science (i.e., scientific methods, scientific knowledge, and its application). Fake news can be defined as “false stories that appear to be news, spread on the internet or using other media, usually created to influence political views or as a joke” (Cambridge English Dictionary online) or simply as “false information.” On the other hand, scientific methods involve principles and procedure for discovering truth in the physical, social and psychological worlds, through systematic observation, measurement, and documentation of reality, as well as formulating, testing, and modifying hypotheses and theories governing the operation of reality.

Certainly, news is not science and may not be judged by the same standards in scientific research but the issue of truth and falsehood concerns both news and science. In fact, some principles of scientific methods are helpful in recognizing fake news particularly when the misleading information is unintended.  This discussion intends to examine five of the principles that help understand false information.

First, opinions or beliefs of a majority do not necessarily represent a source of truth. The truthfulness of a finding or idea does not depend upon how popular it is. It occurs quite often that a new understanding of reality at beginning is dismissed, repressed or denounced.  Take the Einstein Theory of Relativity as an example. Although he developed the theory at the beginning of the 20th century, for a long time, many experts were skeptical about its validity or truthfulness. In 1931, a book was published with the title “100 authors against Einstein,” in the book Einstein's relativity theory was refuted. Most of the authors were leading scientists in their fields or otherwise reputable scholars.

To apply the principle in identifying false information, people need to recognize that even though a piece of news is shared and reported by numerous sources, it does not suggest that the information must be true. They need to look into the supporting evidence and how it is collected and presented.

Second, science studies “what is” or the changing reality of the physical and social human worlds, but does not study “what ought to be” or determine what is morally right or wrong. Making moral judgments of an event, a relation or behavior (i.e., right or wrong evaluations) in research or news are not the same as understanding of its truthfulness or falseness. Just as moral judgments cannot explain biological, physical events or galaxy, moral judgments have little validity in identifying and verifying objective causal, structural, procedural, and other relations among macro- or micro-variables that explain social behavior and psychological activities.  All behavior can be judged from a moral perspective (e.g., right or wrong) and a scientific perspective (true or false), but value appraisals cannot prove what is true or false. In fact, what is true may be morally wrong from the perspective of a perceiver. Of course, all scientists have their own ethical beliefs, which influence how they conduct their research, including what to study, the theory to apply and the methods to collect and interpret the data. However, people may have a strong moral conviction that their understanding of reality is true when their perceptions of reality are incomplete, purely distorted or false. Morally or politically defined correctness is not equivalent to scientifically verified correctness.

One example that shows how people confuse the issues of right/wrong vs. true/false was found in the Medieval Period, when the dominant ideology explaining the universe suggested that the sun rotated around the earth.  In other words, to say the earth rotated around the sun was morally and politically incorrect during that period.

Third, scientists strive in guarding against making generalized conclusions about a large population, group, or category when analyzing observations of some individuals or a small sample of event or behavior. For example, research findings with rats may lack external validity when the results are applied to humans. In the case that a sample is selected from a group, their behavior and attributes may not be representative of the group because people can simultaneously belong to many groups or categories (e.g., education, upbringing, situations, learning experiences, age, ethnicity, and SES) which jointly or differentially shape their target behavior and attributes.

This principle indicates that if a news story hastily generalizes from an individual to a group or category, without justification and eliminating alternative explanations, it tends to be misleading.

Fourth, scientific knowledge is an accumulative and developmental entity. What we know today about the universe, the social world and human behavior is far from complete knowledge, representing far less than 0.00001% of the possible information. Put in another way, there is no truth that is absolute. Science represents a steady process rather than a conclusion. What we know currently may also vary with cultures, history, situations, background of researchers, the processes of observations, and other factors. Consequently, people must continuously discover new reality, engage in innovative research, and revise early theories and knowledge. As Popper (2005) suggested, human knowledge progresses only through the falsification processes (i.e., a theory or idea or its portion being proven false), which distinguish the scientific from the unscientific. The general processes of developing a new scientific concept or model (see Kuhn, 1970) involve comparing both the to-be-rejected model and the to-be-accepted model with reality and comparing the two models with each other. Accepting the new model is based on the discovery that what is anomalous under the old model becomes expected under the new one.

It is certain that a claim or information is false if it declares that the current knowledge about phenomena, events or behavior is complete and should never be challenged.

Fifth, true knowledge or meaningful observations come from comparisons. Experimentation methods or randomized control trials are preferred as the most rigorous way of eliminating bias in discovering truth, because they involve comparing two or more identical groups in terms of attributes except for the experimental treatment.

Let’s look at an example of fake news that has violated the principle. A news media reports that “The product of Company ‘A” has generated 300 complaints because the product is unsafe.” Audience who are exposed to the information are very inclined to avoid the product and develop negative attitudes toward it. What is missing in the news, however, involves the evidence about the number of complaints against and safety records of other comparable companies that manufacture the same type of product.  Another company may have received 500 complaints for its product. Without the information about comparison groups, people’s judgment is easily manipulated.

In short, understanding the principles of scientific methods can help us detect false or misleading information, making more accurate evaluations and decisions about people and events.

References

De keersmaecker, J., & Roets, A. (2017). ‘Fake news’: Incorrect, but hard to correct. The role of cognitive ability on the impact of false information on social impressions. Intelligence, doi:10.1016/j.intell.2017.10.005

Kuhn, T. (1970). The structures of scientific revolutions (2nd ed.). Chicago: University of Chicago Press.

Popper, Karl (2005). The logic of scientific discovery (Taylor & Francis e-Library ed.). London and New York: Routledge/Taylor & Francis e-Library.

More Posts