Abstract
If someone is smiling, it means they're happy, right? Well, not always. Sometimes people smile to be polite, or because they want to "appear" happy or friendly for social reasons. How easy is it to spot which smiles are genuine and which are fake? Try this science fair project to find out!Summary
Objective
Discover whether people can accurately determine if a smile is genuine or fake.
Introduction
How many times a day do you smile? Out of those, how many times do you really mean it? Try keeping track—you might be surprised! People do smile when they're genuinely happy, but the truth is, humans also smile for a lot of other reasons—most of them are social. For example, maybe you've just been introduced to your best friend's cousin, Mark. What do you do? Smile and say hello! But are you really happy to meet him? Probably not; instead you're smiling to be polite and welcoming. On the other hand, if someone introduced you to your favorite movie star—well then you'd be grinning out of sheer delight!
How easy is it to spot a genuine smile—one that happens spontaneously because the person is feeling happy—from one that is made on purpose, for social reasons? Do you think you're good at telling the difference between genuine (real) and fake smiles? Before reading any further, try searching online for "real vs fake smile quiz," "spot the fake smile test," or something similar. Try a few of the tests. How did you do? Are you surprised?
Biologists and psychologists have spent many research hours working to understand how and when different facial expressions are made, and how others interpret those expressions. By observing people who've had strokes or other brain injuries, scientists have learned that there are two parts of the brain that control smiling. The motor cortex controls voluntary motions of the face (that is, motions done consciously and on purpose). When a person wants to smile for social reasons, he or she uses the motor cortex of their brain to do so. But spontaneous, emotionally driven smiles are triggered by a totally different part of the brain: the cingulate cortex. So, as long as the cingulate cortex wasn't damaged, a person who has had a stroke that affects his or her motor cortex can still grin at a good joke if he or she truly finds it funny.
Although the smiles initiated by the brain's motor cortex do a good job of mimicking the emotional smiles initiated by the cingulate cortex, research shows that there are differences between the two types of smiles. The first person to define some of the differences was the nineteenth-century French neurologist Guillaume Duchenne. His work showed that a genuine smile not only stimulated an upward movement of the mouth muscle (the zygomatic major muscle), but also caused a movement in the muscle around the eyes (the orbicularis oculi). The activation of the orbicularis oculi muscle resulted in wrinkling around the eyes (sometimes called crow's feet) and an upward pulling of the cheeks. In contrast, non-emotional or "fake" smiles result in movement of the zygomatic major, but not the orbicularis oculi. In honor of his observations, genuine smiles are often referred to as Duchenne smiles.
Figure 1. The photo on the left depicts a Duchenne smile. Note the uplifted cheeks and wrinkling around the eyes. The non-Duchenne smile, shown on the right, has neither of these features (photo courtesy of Amy and Aneshka Drahota, 2008.)
Researchers in the twentieth century have also noted that in addition to the stimulation of different muscles. Duchenne and non-Duchenne smiles differ in other ways. During Duchenne smiles, the face is generally symmetrical—as opposed to say a smirk, where only one side of the mouth lifts. Duchenne smiles are also smoother in appearance; whereas non-Duchenne smiles often start or end abruptly. And lastly, genuine smiles usually last between 0.5 and 4 seconds; non-Duchenne smiles are often either more fleeting or longer lasting, depending on the social trigger.
So clearly there are physical differences between genuine and fake smiles, but are people innately aware of them? How accurate is the average person at differentiating between a Duchenne and a non-Duchenne smile? You can find out in this science fair project by seeing how your friends and family perform on the Spot the Fake Smile test!
Terms and Concepts
- Motor cortex
- Cingulate cortex
- Guillaume Duchenne
- Zygomatic major
- Obicularis oculi
- Duchenne smile
- Correlation
- Scatter plot
- Best-fit line
- Negative correlation
- Positive correlation
- Pearson correlation coefficient, r
Questions
- How do researchers differentiate between Duchenne and non-Duchenne smiles?
- What experiments led Guillaume Duchenne to conclude that there were two types of smiles?
- What kinds of statistics are used to determine if two variables correlate?
Bibliography
These resources are good places to start learning about smiles.
- Ekman, P. (2003). Darwin, Deception, and Facial Expression. Annals of the New York Academy of Sciences. 1000: 205-221. Retrieved December 19, 2008.
- Stephens, R. (2014, December 22). How we grin to bear it - the science of smiling. Retrieved April 20, 2015.
These websites offer help with creating graphs and more information about correlation.
- National Center for Education Statistics. (n.d.). Create a Graph. Retrieved December 19, 2008.
- Olsen, A. (2006, April 20). Which Team Batting Statistic Predicts Run Production Best?. Retrieved December 19, 2008.
- Shortell, T. (2001). Correlations. An Introduction to Data Analysis & Presentation. Retrieved December 22, 2008.
Materials and Equipment
- Volunteers (at least 40); to find this many people, try asking classmates, friends, family, or sports team members to participate. You will need to be at a computer with each of your volunteers for about 5–10 minutes, one on one.
- Computer with an Internet connection
- Lab notebook
- Graph paper
- Optional: Spreadsheet or statistics program, like Microsoft® Excel®
Experimental Procedure
Testing the Volunteers
For this science fair project, you'll need to test each of your volunteers alone, because you don't want them to be influenced by each other's answers.
- Decide which online real vs. fake smile quiz you will use for your experiment. This project was originally based on the Spot the Fake Smile test, which contained short videos of different smiles, but the test is no longer available online. You will need to choose a different test to use for your project. You can also create your own test by taking pictures or short videos of people smiling. You will have to create situations to get genuine smiles (for example, by making people laugh) as well as asking people to smile for "fake" smiles.
- Ask the volunteer how confident he or she is, on a scale of 1 (low confidence) to 10 (high confidence), that he or she can tell the difference between a non-Duchenne (fake) smile and a Duchenne (genuine) smile. Record the answer in a data table in your lab notebook.
- Have the volunteer use the computer to take the online test. Sit next to them as they work and record their answer (genuine or fake) to each smile. The computer will not display their score, you will have to calculate that yourself. When the volunteer is done, record how well he or she did in the data table in your lab notebook. Note: if your test has videos, make sure the person watches the short video and doesn't just look at the still image of the person's face.
- Record the total number of smiles the volunteer correctly categorized as Duchenne or non-Duchenne.
- Record the number of Duchenne smiles the volunteer misidentified as non-Duchenne.
- Record the number of non-Duchenne smiles the volunteer misidentified as Duchenne.
- After the volunteer has taken the Spot the Fake Smile test and you've shown him or her the score, ask again, on a scale of 1 to 10, how confident he or she is about telling the difference between a non-Duchenne smile and a Duchenne smile. Record the answer in your data table.
- Repeat steps 1-3 for each volunteer.
Analyzing the Data
There are many ways to analyze the data you collected. Different graphs and statistical calculations will help you answer different questions. Below are a few ways of looking at your data to get you started. Try a couple of these, or think of some of your own:
- Calculate the average percentage of correct answers across all your volunteers. How accurate were your volunteers at telling the difference between a non-Duchenne smile and a Duchenne smile?
- Calculate the average number of Duchenne smiles misidentified, and the average number of non-Duchenne smiles misidentified. Were your volunteers more likely to misidentify a non-Duchenne smile or a Duchenne smile?
- Is there a correlation between how confident someone is in his or her ability to distinguish between Duchenne and non-Duchenne smiles and his or her actual ability, as determined by this test? Create a scatter plot to find out. This can be done by hand on graph paper, or on the computer using either a spreadsheet program like Microsoft Excel or a website like Create a Graph. Note: In both Microsoft Excel and Create a Graph, scatter plot is a sub-option found under XY-graphs.
- For each volunteer, place the level of confidence he or she had before the the test on the x-axis and the percentage of his or her correct answers on the y-axis.
- Does it look like there is a trend in the data? If so, draw a best-fit line on your graph. This is a line that best sums up your data.
- If you are drawing this line by hand, try to make it go through the middle of the cloud of data points such that most data points fall evenly on one side or the other. Don't worry if there are some extreme outliers that don't fit the best-fit line.
- If you are using Microsoft Excel, you can have the program add a best-fit line. Use the "Help" features in the program to find out how to do this. Note: Microsoft Excel refers to the best-fit line as a "trend line."
- Does the best-fit line suggest there is a correlation? If so, is it a negative correlation (that is, the two measurements move in opposite directions) or a positive correlation (that is, the two measurements move in the same direction). If there is a correlation, would you describe it as strong or weak? Note: To learn more about the strength of correlations, look at the references in the Bibliography.
- Optional: The correlation between two variables can be described by statistics. Advanced students should calculate the Pearson correlation coefficient, r, for their data. If you need help with this calculation, the Which Team Batting Statistic Predicts Run Production Best? science fair project has a step-by-step introduction to correlation and linear regression. Note: Excel will calculate r2and the slope of the trend line; you can back-calculate r from r2.
- Does taking the Spot the Fake Smile test change people's confidence in their ability to detect Duchenne versus non-Duchenne smiles?
- For each volunteer, calculate the change in confidence level by subtracting his or her post-test confidence level from his or her pre-test confidence level.
- The number will be positive if the test increased the volunteer's confidence, 0 if the test did not alter confidence, and negative if it decreased his or her confidence.
- Create a pie chart showing the percentage of volunteers whose confidence levels were unchanged, increased, and decreased. What was the most common effect of the test on confidence? Can you relate this to your observations about how accurate people are at telling the difference between non-Duchenne and Duchenne smiles?
- For each volunteer, calculate the change in confidence level by subtracting his or her post-test confidence level from his or her pre-test confidence level.
Ask an Expert
Variations
- Does outlook on life (pessimism versus optimism) have an effect on people's ability to detect non-Duchenne smiles? How about age or gender? Or personality traits like extroversion versus introversion? Design experiments to find out.
- Does knowing about Duchenne versus non-Duchenne smiles change people's ability to detect genuine versus fake smiles? Try modifying the experimental procedure above to find out. Hint: You'll need to teach one group of people about Duchenne smiles, but not the other. Both test groups should consist of at least 40 volunteers each, to increase the confidence in your results.
- Can people "hear" a genuine smile in someone's voice? You'll need to devise a way of having people say the same sentence repeatedly while sometimes having a Duchenne smile and sometimes not. One method might be to video tape people reciting a funny poem with a refrain. Use the video tape to catalog when they have a Duchenne smile on their faces and when they don't. Then play audio clips associated with Duchenne and non-Duchenne smiles (or no smile at all) for other volunteers—can they distinguish, by voice alone, when someone is genuinely smiling?
Careers
If you like this project, you might enjoy exploring these related careers: