27 December 2019

Biased Memories

Schema (pl. schemata)…mental structures that an individual uses to organize knowledge and guide cognitive processes and behaviour…Schemata represent the ways in which the characteristics of certain events or objects are recalled, as determined by one’s self-knowledge and cultural-political background. (Britannica)

Welcome back. There’s a nice, neat, very cool study by Ohio State University researchers on false information, specifically numbers. It turns out that people misremember numerical statistics to fit their schemata, their beliefs or expectations. To show this, the researchers conducted two experiments.

Misremembering Unexpected Numbers
In the first, they had 110 participants read short descriptions about four issues that contained numerical information. Pre-testing indicated that the numerical information on two of the issues would fit most participants’ expectations, while the numerical information on the other two issues would not.

An example of the former is that people generally expect more Americans to support rather than oppose same-sex marriage, which is true. An example of the latter is that most people believe the number of Mexican immigrants in the U.S. increased between 2007 and 2014, though it actually decreased by 1.1 million.

After reading the four descriptions, the participants were given a pop quiz: Write down the numbers in the descriptions.

Misremembering numbers that
don’t fit expectation.
The participants usually got the numerical relationship correct on the two issues that were consistent with how most people viewed the issue, such as same-sex marriage. But on the two issues that did not fit their expectation, such as the trend in Mexican immigrants, participants were much more likely to remember the numbers according to their probable biases rather than the truth. Some remembered the exact numbers, reversed.

As an extension to the first experiment, the researchers used eye-tracking technology to monitor participants when they read the descriptions. The eye responses were very different when reading numerical information that fit and didn’t fit their expectation.

It Gets Worse When Spread
For the second experiment, the researchers tested how social transmission of numerical information can exacerbate the memory errors.

The Telephone Game (photo from:
Similar to the Telephone Game, the first participant in a “telephone chain” read the correct statistics about the trend in Mexican immigrants in the U.S. That person wrote down the numbers from memory, and the numbers were passed to the second participant in the chain. That person wrote down the numbers from memory, the numbers were passed to the third participant, and so on down the chain.

The experiment found that, on average, the first participant reversed the numbers, remembering an increase of 900,000 immigrants instead of a decrease of 1.1 million. That error increased to about 4.6 million immigrants by the last participant.

Wrap Up
The study showed that self-generated, numerical misinformation can be as bad as, if not worse than, false information from external sources. Unfortunately, our biases are constant.

The study was published about a week before the Washington Post Fact Checker reported that President Trump made 15,413 false or misleading claims over 1,055 days. Seeing the updated Fact Checker statistics, I wondered if, instead of intentionally stating numerical information wrong, the president was just misremembering to fit his beliefs. Probably not considering the preponderance of non-numerical claims.

Thanks for stopping by.

Study of numerical misinformation in Human Communication Research journal: academic.oup.com/hcr/advance-article-abstract/doi/10.1093/hcr/hqz012/5652186?redirectedFrom=fulltext
Ohio State University news release on study: news.osu.edu/you-create-your-own-false-information-study-finds/
Washington Post article on president’s false or misleading claims: www.washingtonpost.com/politics/2019/12/16/president-trump-has-made-false-or-misleading-claims-over-days/

No comments:

Post a Comment