Purpose: Students will learn how to conduct research and design scientific research studies.
Procedural overview: After the teacher introduces the concept of face pareidolia, students will photograph faces they perceive in everyday objects before creating a class poll to determine whether the faces are perceived as male, female or neutral. After results of the class poll are tabulated, students will read and analyze the article “Americans tend to assume imaginary faces are male” from Science News and compare their results with those from the research study presented in the article. Students can then design a follow-up study that expands on the research.
Learning outcomes: Asking questions and defining problems; constructing explanations and designing solutions; engaging in argument from evidence; obtaining, evaluating, and communicating information
Approximate class time: 1 to 2 class periods
Cell phones with cameras
Directions for teachers:
To prevent bias in results, do not assign the Science News article before the start of this activity.
Use the following prompt or one you create to introduce the activity.
Have you ever seen a cloud that looks like a dog’s face? Or eyes, a nose and a mouth in a loaf of bread? Well, it isn’t just you! Most people experience this phenomenon, which is called face pareidolia. It is our brain’s way of making something familiar and meaningful out of random or unclear visual information.
To prepare for the activity, ask students to take photographs of faces they perceive in nature or in inanimate objects. Consider having students work in pairs or small groups because some students might see faces more readily than others.
Have the class take at least 20 and as many as 100 photographs. A wide variety of photos could lead to more interesting results. Compile the images that are collected into a shared file that students can access when it is time to conduct the class poll.
Want to make it a virtual lesson?
If students are working remotely, they can do the following categorizing using a conferencing platform such as Zoom. Zoom also has a polling feature that would allow students to submit their answers digitally and a virtual whiteboard to display the results of the poll.
Categorizing the images
Decide how many images you want to use for the class poll. Number each image and offer female, male and neutral answer choices for each image.
After students view the images and take the poll, have them tally for each image the number of students who viewed the image as male, female or neutral; also calculate the percentages. Then do a tally for the whole collection of images to determine how many were viewed as male, female or neutral. Again, calculate the percentages. Ask students to summarize the results.
After students complete the poll and analyze the results, hand out the student worksheet and explain that students will be learning about scientists who did a project like theirs.
Ask students to read the online Science News article “Americans tend to assume imaginary faces are male,” which describes researchers’ investigation into how humans characterize faces in inanimate objects. A version of the article under the same title appears in the March 12, 2022 issue of Science News.
Use the following questions to analyze the article.
1. What is a bias?
A tendency to prefer something over other options, either intentionally or unintentionally.
2. What do you think the scientists’ hypothesized?
Student answers will vary. The scientists might have hypothesized that people tend to see male faces in inanimate objects. Or, they could have hypothesized that people are just as likely to perceive a female face as a male face in an inanimate object.
3. What was the purpose of repeating the experiment with images of objects that had no faces?
The purpose of repeating the experiment with images of objects with no faces was to see if the objects themselves contained feminine or masculine traits that might have biased people’s perceptions when they saw the images of the objects.
4. What were the results of the experiment?
When looking at illusory faces in the experiment, people perceived males faces four times as often as female faces. 80 percent of participants identified over half of the faces as male; 3 percent of participants identified over half of the faces as female, and 17 percent of participants were even in their labeling.
5. How did the class’s results compare with the results of the research experiment in the article?
Student answers will vary. Both our class and the scientists found that illusory faces were most often perceived as male.
6. If the class’s and the scientists’ results are very different, what might explain the difference?
There could be a difference in how the experiments were run. For example, our poll might have given away the potential bias we were looking for. The age of the participants might have differed as well. Our class is only one grade level/one age group and the scientists polled adults. We also used a different number of photos, and we used different types of photos than the ones in the scientists’ experiment.
7. Why do you think that people might be biased toward seeing male faces?
People’s bias to see male faces could be related to their cultural and social experiences — for instance, how they were raised.
Designing the next study
After the students read and analyze the Science News article, they will design a follow-up research investigation.
If necessary, review research design with your students. Terms such as independent variable, dependent variable and testable hypothesis are covered in Cookieology: Experimental Design 101.
Have the students use the following questions to guide their hypothesis-making and experimental design.
1. Based on the research results, what questions do you have about face pareidolia?
Student answers will vary. Sample answers include: Are males more likely to see male faces than female faces in inanimate objects? Do people see young faces more than old faces? Are there people who don’t see faces in inanimate objects?
2. Pick a question that you want to address and write a testable hypothesis.
Student answers will vary. I want to see if high school students are more likely to perceive faces as young than as old. I will define young as under 30. My hypothesis is people in high school are more likely to imagine they see a young face than an old face when they see faces in inanimate objects.
3. What can you do to test your hypothesis? Explain the design of your research experiment.
Student answers will vary. I will do a poll like the one the class did to determine whether people perceived young faces more than old faces. I will mark the images with young, old and neutral; then I will ask my classmates to rate them. Then I will poll the whole class.
4. What are your variables?
The images of objects with faces and the participants’ opinions about whether they look young or old are the variables.
5. What materials will you need to run your experiment?
Student answers will vary. I will need computer access, a way to mark the images and a way to survey the class.
6. How will you collect data?
I will tally all responses provided by individual participants and then do a final tally for the whole class’s results.
7. When people are aware of a bias, they are likely to change their answers or responses. For example, someone who knows that people are biased to see male faces might avoid choosing male as an option because they know about the bias. How might you choose your participants to make sure you are getting reliable answers?
I will gather data from other students who are not in my class during lunch.
Activity extension: Conducting the next study
After designing their research study, students can run the study as an activity extension. The photographs the students took can be used as a resource when creating any new polls, which can be done by surveying another class or outside of school. After students have collected their data and summarized their results, have them answer the following questions.
1. What challenges did you have while running your experiment?
Student answers may vary. I had a hard time getting people to do the survey during school lunch and outside of school.
2. What were your results?
Student answers will vary. My results were not clear. When I surveyed people outside of school, they were a variety of ages. There was no pattern in who was biased toward young faces. But young people at school seemed to view the images as young.
3. Did the results support your hypothesis? Why or why not?
Student answers will vary, but should provide the reasoning behind their hypothesis. No, but I think the problem was in how I did the study and who was in it.
4. If you were to do the same experiment again, what would you do differently?
Student answers will vary. If I were to do this experiment again, I would like to ask people over 60 to take the survey. I am curious whether their perception of young and old faces would be the same as the results for teenagers.