Different types of smiles

Challenge

This project did not specifically have a user.
The intention was to broaden our knowledge about how individuals perceive other individuals
A research team found out that if a person smiles in a way that matches how you ideally like to feel, you tend to like that person more (Tsai, et al., 2019)
The two ways people ideally like to feel positive emotions (Tsai, 2006):
- High arousal (excitement): many people prefer to experience excitement, elation, and other types of high arousal positive emotions
- Low arousal (calm): many people prefer to experience calm, peace, and other types of low arousal positive emotions
The research team found that if you prefer:
- Excited emotions, you will like a person more if they have an excited smile
- Calm emotions, you will like a person more if they have a calm smile
This team found that the race of the other person did not affect this judgment
- However, they were only looking at Caucasian and East Asian faces.

Goal

Examine if people like others with matching emotional expression across other races.

Why is this project important?

Positive relations between different groups is key to a country's success, especially in such a country as diverse as the U.S.

Understanding how a person's smile affects the way someone of a different race views them adds to our growing knowledge of intergroup relations and contributes toward advancing positive intergroup relations.

Result

We found that White American participants preferred White individuals with excited smiles
As we expected, there was no difference in how much White Americans liked excitedly smiling Black individuals versus calmly smiling Black individuals
- This result may be due to participants not wishing to rate Black individuals negatively and needs further exploration
Contrary to previous research, there was also no difference in how much White Americans liked excitedly smiling Asian individuals compared to calmly smiling Asian individuals.

Team

My team consisted of:

My supervisor, Dr. Diane Mackie - Distinguished Professor in the Psychological and Brain Sciences (PBS) Department at the University of California, Santa Barbara (UCSB)
Dr. Kyle Ratner - Assistant Professor in the Psychological and Brain Sciences (PBS) Department at the University of California, Santa Barbara (UCSB)
Dr. Heejung Kim - Professor in the Psychological and Brain Sciences (PBS) Department at the University of California, Santa Barbara (UCSB)
Vinnie Wu – Graduate student in the Psychological and Brain Sciences (PBS) Department at the University of California, Santa Barbara (UCSB)
Youngki Hong – Graduate student in the Psychological and Brain Sciences (PBS) Department at the University of California, Santa Barbara (UCSB)
Rammy Salem – Graduate student in the Psychological and Brain Sciences (PBS) Department at the University of California, Santa Barbara (UCSB)
Sierra Feasel – Graduate student in the Psychological and Brain Sciences (PBS) Department at the University of California, Santa Barbara (UCSB)
And my Research Assistants who are all undergraduates in PBS at UCSB
1. Courtney Chan
2. Jessamy Johnson
3. Justine Johnson-Yurchak
4. Katie Sabini

Research process

Stage 1: Identifying the problem

I attended a conference where I heard the above referenced research team discuss their findings. At the same conference, I learned that another team had conducted research involving Black targets and found that these targets were treated differently from Caucasian and East Asian targets. However, the latter group had only done work on body postures and not smiles. Thus, I identified a gap in our understanding of perception and how race and simple behaviors like smiles can affect these perceptions.

I conducted a literature review to ensure that this was indeed a gap in our knowledge and looked at other work that had been done on smiles as well as why different groups of people may react differently to people of different races.

Given my literature review, I formulated the following hypothesis:

Caucasian Americans (who research shows to prefer excited smiles) will rate East Asian and Caucasian individuals with excited smiles (compared to calm smiles) more positively, but will not show a difference in their ratings between Black individuals with excited compared to calm smiles.

I then presented my literature review and hypothesis in an in-person presentation to my team and gathered feedback on the type of study I should run to test this gap in the literature.

My team agreed to run a study with a mixed design. Caucasian American participants would view East Asian, Black, and Caucasian individuals’ faces and rate them on several traits related to extraversion and agreeableness. Participants would either see only excited smiles on these faces or only calm smiles.

Stage 2: Designing a study to address the problem

Given that we were to present a set of Black, East Asian, and Caucasian faces to our participants, I searched for, requested access to, and inspected several face databases.

I finally decided on using the Chicago Face Database (Ma, Correll, & Wittenbrink, 2015) for the Caucasian and Black faces, and the Taiwanese Facial Expression Image Database (Chen & Yen, 2007) for the East Asian faces.

There were differences in the faces between the two databases and we had to engage in photo editing to ensure that all our material was similar.

I assisted my teammate, Courtney Chan, with finding a photo editing software, and together with my other teammate, Dr. Diane Mackie, we worked on creating face images that were as similar to each other as possible.

With the help of the other Research Assistants, I pilot tested my materials, by asking them to rate these faces on several traits including attractiveness and trustworthiness.

After the materials had been pilot tested, I designed a survey on Qualtrics in which participants would be exposed to either excited or calm faces, and then view each of the faces in their condition randomly and rate them on a set of traits.

Stage 3: Obtaining data

Once the survey had been completed and my teammates and I had run through the survey several times, we launched our experiment.

At first, we collected Caucasian American participants from the online participant recruitment platform, Amazon MTurk.

However, we ran into a problem as our data collection was advancing slower than we anticipated. To speed up the process, we made changes to several aspects of the recruitment process without changing the main principle that our participants were Caucasian Americans. We also opened up recruitment in our university and recruited Caucasian American UCSB students.

Thus, our data collection took place through both an online and in-lab survey study.

Stage 4: Understanding the data

All data cleaning, manipulation, visualization, and analysis was conducted in R (although since then I have re-run some of my data wrangling through SQL as well).

The first step after obtaining the data was to engage in data cleaning. First, I renamed a few of my data columns to make the database more user-friendly. Next, I focused on weeding out responses that would muddy the data. I had added a few attention check items into my survey to ensure that participants were actually reading the prompts that they were given. I wrote a script to identify participants who had made 1+ errors on the attention check items and removed their data.

Next, I ran data manipulation to change the format of the data such that I would be able to run the necessary statistical analyses on it. I first created the composites for extraversion and agreeableness by averaging across the items meant to measure these two forms of positivity. Each participant received three extraversion and three agreeableness scores, one for each race.

I then created a column that designated whether each participant had seen excited or calm smiles and a column that designated whether the participant was a UCSB student or an Amazon MTurk worker.

I next turned my data from wide format to long format. An example of what the final dataframe looked like is shown below (the data is hidden):

Next, I ran a series of mixed ANOVAs to see whether the race of the individual and whether they had an excited versus calm smile would predict the extent to which participants found this individual agreeable or extraverted.

The ANOVAs indicated that only part of my hypothesis was correct:

As expected, Caucasian American participants rated Caucasian faces with excited smiles more extraverted and agreeable than those with calm smiles. They also showed no difference in their ratings of Black individuals with excited compared to calm smiles. However, these participants also did not show a difference in their ratings between East Asian faces with excited and calm smiles.

To explore potential reasons for these results, my team had also decided to ask participants to rate these faces on aggressiveness and foreignness.

We wanted to explore whether perceived aggression or foreignness of faces drove the results in a particular way.

So, I conducted several mediation analyses in which I looked at whether calm smiles (the less preferred smile) was rated as more aggressive or foreign than excited smiles (the more preferred smile) within each race.

I found that participants rated Caucasian faces with excited smiles as significantly less aggressive than Caucasian faces with calm smiles and that this tendency predicted their liking of Caucasian faces with excited smiles over Caucasian faces with calm smiles.

I wrote up all my steps, including data cleaning, data manipulation, ANOVAs, and mediation analyses on an R Markdown document, knit it into a Word document, and sent it to my teammates. This allowed for teammates who were unfamiliar with R to be able to read through my steps and results and provide me with feedback.

Next, I engaged in broader forms of advocating for my research. I presented these findings in two lab meetings, helped my teammate, Courtney Chan, create a poster of this study, and will be presenting this study to an international conference in February, 2021, where I will discuss my results with both experts and novices in the area.

Different types of smiles

CONTACT