Accuracy: Are You Only Swimming in the Shallow End?

| By TMA World

Do we tend to overestimate our ability to apply our existing knowledge accurately?  Psychologists demonstrate this dysfunctional tendency by having us take simple tests, such as:

  • In the biblical story, what was Jonah swallowed by?
  • How many animals of each kind did Moses take on the Ark?

If you have any knowledge of these biblical stories (they are obviously not culturally neutral), you probably answered “whale” to the first question, and “two” to the second.   While “two” is arithmetically correct, the second question contains a major factual error.  It was Noah, and not Moses, who took animals “two-by-two” onto the Ark.  You probably knew that already.

This psychological phenomenon has come to be known as the “Moses Illusion” (Why you are so bad at fact-checking those April Fools Day pranks in The Washington Post, April 1, 2018).  In the original study, 80 percent of participants failed to notice the error in the question although later they correctly answered the question “Who was it who took the animals on the Ark?”  Surprisingly, participants failed the initial question even though they were warned prior to the test that some of the questions would have something wrong with them.

As well as the “Moses Illusion”, this phenomenon is also called the semantic illusion or knowledge neglect.

The problem with knowledge neglect is that it can lead us down a path toward more significant and damaging errors.  

Why Does This Happen?

  • Plausibility: Plausibility of the incorrect information is one explanation for the error.  Moses fits the Old Testament context, and at a glance is highly plausible.  People will often “go with the flow” and accept presented information as “good enough.”  It takes a cognitive effort to spot Moses Illusions! 
  • Language Processing: Psychologists used to think that we understand a sentence by weighing the meaning of each word, but further research suggests this isn’t true.  Instead, our language processing is based on shallow and incomplete interpretations of what we hear and read.  Here is a well-known example: “After a plane crash, where should the survivors be buried?”  The brain seems to pick up on words like “crash” and “buried”, and roughly half the people in studies answer the question as if they are being asked about the victims and not the survivors.  And consider this: “Can a man marry his widow’s sister?”  Most people answer “Yes”, but are they really agreeing that a dead man can marry his bereaved wife’s sister. 
  • Context: If I encounter a series of statements or questions that are accurate, and then one that contains a distortion, I know that my expectation of accuracy in any subsequent statements will be stronger.  When a reader encounters more errors in a sequence, it seems to lead a reader to evaluate each individual question more thoroughly.  

Does Anything Help?

  • Processing Obstacles: In one interesting finding, error detection improves when questions appear in difficult-to-read-fonts; this reduces processing fluency which makes material seem less familiar and less true.
  • Knowledge: Increasing knowledge of the topic seems to help, but doesn’t solve the problem.  It has been shown that even biology graduate students will attempt to answer distorted questions like, “Water contains two atoms of helium and how many atoms of oxygen?”  Water, of course, contains two atoms of hydrogen, not helium. 
  • Highlighting: Researchers have tried highlighting critical information in red font.  It was hoped that the red highlighting would enable participants to more easily notice errors and avoid them on a later test.  That didn’t happen. 
  • Time: Are people more likely to notice errors if they have more time to process the information.  Researchers have presented the information in a books-on-tape format and slowed down the presentation rate.  Again, the desired result didn’t happen.
  • Role: One thing that seems to help is having participants in research studies act like professional fact-checkers.  Participants have been asked to edit information and highlight any inaccuracies.  As a result, participants are less likely to learn misinformation.  A similar result has been achieved when participants have been asked to spot errors by read and evaluate information sentence-by-sentence.

We cannot, however, get complacent.  In the sentence-by-sentence detection task participants caught about 30 percent of the errors.  Given their prior knowledge they should have been able to detect at least 70 percent.

Swimming in the shallow end of knowledge and language processing can be expensive and even devastating.  We must remember that in pursuing accurate understanding we are not alone.  Hell might be other people, as Jean Paul Sartre said, but they can also be our salvation.  Let me see the figurative crowd wave for “Collaboration!”


Are your organizational L&D programs having the desired impact on productivity and efficiency?

At TMA World we specialize in delivering learning programs for complex organizations working across borders – be that hierarchical, geographical, cultural or technological. We develop the skills and capabilities needed to thrive in global organizations – where leaders, managers and individual contributors are navigating matrix structures, cultural differences, remote working and virtual technologies. Our content is available in various formats, bespoke to our client’s individual needs and requirements. Click here to view our success stories.

If your organization would like to develop a truly borderless workforce, we would love to hear from you. Click here to speak to one of our team.  

Interested in how introducing a cultural intelligence tool in your business could help to create a more borderless workforce? We’d love to show you our groundbreaking platform.