Confirmation Bias, Politics, Research and Last-Minute Studying: Individuals have lots of assumptions about the way the world works, assumptions that help frame how they make decisions. These assumptions are often challenged or confirmed by empirical evidence. However, psychologists who have analyzed how people change their assumptions about the world suggest that we tend to seek out evidence that confirms our assumptions and ignore evidence that contradicts our assumptions. This phenomenon is known as confirmation bias, and one of the early experiments uncovering this bias is described in part B.
A: Over the past few decades, there has been a vast increase in the number of sources that individuals can use to inform themselves about what is going on in the world. For instance, most individuals used to rely on their local newspaper (which often drew its material primarily from a handful of national news outlets) and the evening newscast on one of three networks. Today, on the other hand, there are lots of cable news channels people can choose from throughout the day, and an increasing number of people rely on news from internet sources.
(a) Many observers of public discourse have suggested that the assumptions individuals bring to policy discussions are now often more diametrically opposed than in the past, with different camps often no longer able to hold civil dialogue because they so fundamentally disagree about the underlying “facts”. If this is true, how can this are explained by the increased number of news and opinion outlets?
(b) In the past, opinion polls often suggested that public disapproval of a U.S. President was in the single digits, but more recently, a President is considered as doing well if his disapproval ratings are in the 20 to 30 percent range. Can confirmation bias in the more recent news environment explain this?
(c) Until the mid-1980’s, the Federal Communication Commission in the U.S. enforced a rule known as the “Fairness Doctrine”. This rule required news outlets — particularly on radio and TV—to present opposing viewpoints. It was argued at the time that some media markets only had one or two such news outlets, and thus the Fairness Doctrine was required to allow people to get alternative points of view so that they could then form in formed opinions. Since themid-1980’s, the Fairness Doctrine is no longer applied—allowing news outlets to present news and opinions in any way they see fit. It was argued that increased competition has led to competing news outlets in virtually all markets — thus automatically allowing individuals to gather alternative viewpoints to form their own opinions. Now some are arguing for a reinstatement of the Fairness Doctrine but others view it as a violation of free speech and free competition of ideas in the product-differentiated marketplace. Can you argue both sides of this issue?
(d) Some have observed an increase in the number of people who believe in a variety of “conspiracy theories” — theories such as that the 9/11 attack was orchestrated by the government or that a politician secretly adheres to a religious view that differs from his stated view. How might this be explained in light of the fact that most individuals find evidence against such theories conclusive?
(e) Empirical social scientists often do econometric regression analysis on real-world data to ascertain the direction and magnitude of people’s responses to different policies. As computational analysis has become less costly, such researchers are now able to run literally tens of thousands of different regressions — using combinations of different variables and empirical specifications — whereas in the past they have had to limit themselves to a few regressions. Suppose that researchers have prior beliefs about what an empirical investigation might show. How might you view statistically significant empirical results reported in research papers more skeptically as a result of knowing about confirmation bias?
(f) In the final hours before an exam, students often “study” intensely by scanning their notes and focusing on key terms that they have highlighted. Some students find that this dramatically increases their sense of being prepared for the exam—and then find that they do not do nearly as well on the exam as they had thought they would given their last-minute studying. Can you explain this using the idea of confirmation bias?
B: The following experiment, first conducted in the early 1960’s, is an illustration of confirmation bias. Suppose that you are given the following sequence of numbers: 2-4-6. You are told that this sequence conforms with a particular rule that was used to generate the sequence and are asked to figure out what the underlying rule is. To do so, you can generate your own 3-number sequences and ask the experimenter for feedback on whether your sequence also conforms to the underlying rule. You can do this as often as you need to until you are certain you know what the underlying rule is— at which time you tell the experimenter your conclusion.
(a) Suppose that, when you first see the 2-4-6 sequence, you recognize it as a sequence of even numbers and believe that the underlying rule probably requires the even numbers. What is an example of a sequence that you might use to test this assumption if you have confirmation bias?
(b) What sequence of numbers might you propose to test your assumption if you did not have confirmation bias and were open to your assumption being incorrect?
(c) The underlying rule was simple: in order to comply with the rule, it simply had to be an ascending sequence. Very few subjects correctly identified this rule — instead very confidently concluding that the rule was much more complex. The experimenters concluded that people consistently derived an incorrect rule because they gave examples that would confirm their assumptions rather than attempt to falsify them. (A sequence intended to “falsify” an assumption would be one that violates the assumption.) How is this consistent with your answers to (a) and (b)?