Psychology Digest

Confirmation bias influences how we process information. When confronted with new data, we tend to favor information that aligns with our existing beliefs. Confirmation bias is more likely to occur when we are confronted with information that challenges beliefs that are important to us. In these situations, we are more likely to accept flawed arguments that are consistent with our existing beliefs and to dismiss contradictory information even if it is strong. Political beliefs provide a good example. Democrats may dismiss arguments that Trump was right about the problem of trade rules with China, whereas Republicans may ignore evidence that Trump actively tried to overturn the outcome of the 2020 election. Confirmation bias creates division between people because they disagree about the truth of their beliefs.

Confirmation bias starts with the search for information. People tend to search for information in places that supports their beliefs. This may happen because they are surrounded by like-minded people or because they actively seek information from sources that confirm their beliefs (e.g. watching FoxNews or CNN).

Confirmation bias can also influence the interpretation of information. Biden and Trump have made mistakes in their speeches, but these mistakes are interpreted differently by Democrats and Republicans as a symptom of cognitive decline.

Finally, confirmation bias can also influence memory. We are more likely to recall consistent than inconsistent information with our beliefs.

In theory, it might seem easy to avoid confirmation bias. For example, people could see how different news outlets cover the same topic. In practice, there are several reasons why confirmation bias is so common. First, people may lack motivation to look for inconvenient truths that challenge their world view. For example, people who just bought a car running on gasoline are unlikely to read about the advantages of electric vehicles. Second, people may lack time or other resources to search for information. For example, consumers may not want to pay for information on consumer reports and rely on biased online reviews that are free.

Evidence

Ironically, evidence in support of confirmation bias may be biased by confirmation bias. Scientists are human and are also prone to confirmation biases (Nickerson, 1998). Psychology articles are much more likely to report the confirmation of a hypothesis than disconfirming evidence. This bias is known as publication bias, but the underlying cause is confirmation bias by a scientist or a group of scientists who share a common belief. Publication bias led to the replication crisis in social psychology, where only 25% of published results could be replicated when independent researchers without confirmation bias tried to reproduce the original results ( Open Science Collaboration, 2015). Thus, the question is whether evidence for confirmation biases is replicable or not.

A major literature review (Nickerson, 1998) and a meta-analysis (Hart et al., 2009) were published before the replication crisis questioned the validity of published results. In fact, one of the studies in the meta-analyses (Fischer et al., 2008) failed to replicate in the Reproducibility Project (Open Science Collaboration, 2015). It is therefore necessary to look for new, credible evidence to provide scientific evidence of conformation bias and to quantify the magnitude of this bias.

One study (Frimer, Skitka, & Motyl (2017) provided evidence of confirmation bias in a blatant choice-task that asked participants whether they would like to hear arguments from like-minded or opposing (e.g., Obama vs. Romney voters) people. Results across several studies provided consistent and strong evidence that participants favored to hear from like-minded people with moderate effect sizes. The moderate effect size implies that people differ in the extent of confirmation bias and that some people would rather be interested in hearing from the other side.

Minson and Dorison (2022) review evidence that people feel unpleasant when they are exposed to opposing views. This can explain a preference for information that is consistent with one’s views. Minson and Dorison go beyond this evidence by distinguishing different types of unpleasant feelings. Insecure people might react with anxiety when they are opposed to strong arguments that challenge their views. In contrast, confident people are likely to respond with anger.  Their results show that more people respond with anger than with fear. This suggests that overconfidence contributes to confirmation bias. It also explains why experts are often the least likely to be aware of biases because their status as experts makes it ignore others that are deemed less competent.

Minson and Dorison (2022) also speculate that overconfidence may foster confirmation bias in another way. Processing information takes time and resources. A cost-benefit analysis implies that experts are less likely to benefit from reading about a topic because they already are familiar with most of the arguments. Therefore, they may not recognize that they could still learn from new information to correct biases in their own thinking.

While evidence for the existence of confirmation bias is strong, it is also clear that the magnitude of confirmation bias is influenced by many situational and personality factors. However, research on personality differences in susceptibility to confirmation biases is rare (Berthet, 2021). If confirmation bias is often rooted in overconfidence, the literature on personality and overconfidence might provide some insights, but more research on personality and confirmation bias is needed.