In some high school psychology classes, students are shown a video that demonstrates selective attention. This video shows two teams bassing basketballs to each other, and students are instructed to count the number of passes for one of the teams.

The teacher asks for their answers, the students count about about 12-15. Then the teacher asks, “but did you see the gorilla?”

Rewinding the video reveals that there is a person wearing a gorilla suit, walking in plan views through the middle of the scene, which bewilders most of the students – how did they not notice that?!

I also read an amusing article on testing the capability of LLMs for explorative data science by Chiraag Gohel, which talks about a study. by Itai Yanai and Martin Lercher published in 2020 that tested undergraduates ability to see the gorilla when working with a fictitious dataset. Hilariously, when the data is plotted on a scatter plot, the picture of a gorilla appears, which obviously does not appear when over analysing the statistics of the dataset. Students tasked with investigating a single hypothesis were apparently 5 times less likely to notice the gorilla that students who were tasked with an open-ended investigation. This raises concerns about hypothesis-driven research as a whole

This is a unifying feeling, at some stage we’ve experienced something like this. I’ve personally experienced it several times in mathematics and programming. When you’re so focussed on solving a problem one way, you have to learn to actively look out for other clues that may reveal alternative successful solutions. I don’t think there’s a reliable solution to this problem when working in isolation – the focus needed to solve a problem one way takes away from the focus needed to keep the scope broad.

I’ve been thinking about how this affects life more broadly. A post by the Association for Psychological Science explores this in more detail, and considers the life-impacting consequences of “inattentional blindness” in the medical field

There’s three lessons I can identify from these anecdotes:

1. Awareness
2. Exploration
3. Collaboration

Being aware of the problem gives us pause, and can give us opportunity to stop and survey where we are to make sure we’re not missing something crucial

Taking a exploratory approach instead of a hypothesis-driven approach also helps tremendously. When I have a goal of solving a problem, I first focus on learning as much as I can about the problem and planning different methods to solutions before jumping into the rabbit hole

We all have different perspectives and blind spots. We are inherently a social species, and historically I think we’ve overcome many gorillas in the data over centuries by working together.

This lesson seems increasingly more important recently. There are no shortage of spectacles in the news, ridiculous events and actions meant to take advantage of this quirk in our attention. These spectacles often force us collectively into a hypothesis-driven view, making it easier for the gorilla to slip through

We can try and find ways to tackle this with exploring the data more often – but there is no substitute for collaboration. Together we can work together to overcome this quirk of psychology. As we make more progress on large problems together with our friends and neighbours and colleagues, we can look out for each other. When something seems off, and we identify blind spot for one or several people, we can ask:

Did you see the gorilla?