Can Your Personality Predict Your Favorite Flavors?

We found that someone’s personality can help explain their taste preferences. Using Truity’s Decision Style Test, we surveyed over 6,000 people and found a striking association between personality and taste. Like this tasty insight: people whose favorite foods are sweet are significantly more likely to be ‘sweet’ towards others! Applying machine learning (ML) tools to the data allows us to guess people’s favorite flavor with approximately 50% accuracy given only their personality results—random guessing would be correct only 20% of the time.  

Introduction: taste and personality

In English, the word taste is known to mean a couple of things. It can refer to someone’s decision-making style (e.g., “she has no taste for negativity”) and to their flavor preferences (e.g., “this cake tastes great!”). We also use “taste” words like sweet, sour, and bitter to describe people’s moods, personalities and behavior patterns. 

This is not only true for English. For instance, in Spanish, the verb gustar both means "to want" and "to taste."  In Italian, gustare means "to enjoy" and "to taste." In German, the word Geschmack also refers both to liking and to tasting.

Are these linguistic coincidences, or is there a legitimate psychological relationship between our two types of tastes? For example, could people who are more sensitive to positive emotions also be more responsive to the dopamine release associated with "sugar highs?" Are "sweet" people more likely to prefer "sweet" foods? 

Following up on related investigations, we asked these kinds of questions as a research project—and the short answer is yes, in general, our personalities are statistically related to our favorite foods and flavors!

Humans have five major taste receptors on our tongues—sweet, salty, umami (savory), sour and bitter. Our goal was to understand whether people’s relative flavor preferences (e.g., “would you prefer a sour or a salty snack?”) could be predicted using their relative psychological preferences (e.g., “would you prefer to have a career in a competitive firm or doing important social work?”). 

Using data collected from over 6,000 responses to Truity’s Decision Style Test, we analyzed whether participants’ personality results could predict their taste profiles. Below is a high-level summary of the findings. 

Q1: Salty vs. sour flavor preference

Key findings:

• People who are more sensitive to negative emotion (e.g., cautious, higher anxiety, higher volatility) are very significantly more likely to prefer salty foods over sour foods. (p≈0, t=6.2686).
• People who are more likely to generally prioritize the wants and needs other people over their own are very significantly more likely to prefer salty foods over sour foods. (p≈0, t=4.1614).
• People who are more interested in literature, art and music than science, technology and logical reasoning are significantly more likely to prefer salty foods over sour foods. (p=0.0027, t=3.0005).
• People who are more sensitive to positive emotion (e.g., are more bubbly, enthusiastic, assertive) are significantly more likely to prefer sour foods over salty foods. (p=0.0013, t=-3.2229). 
• People who are more likely to prioritize their own learning and growth over other people are significantly more likely to prefer sour foods over salty foods. (p=0.0447, t=-2.0079). 

How accurate is the model? 

Given participants’ total trait scores, the trained model can correctly predict whether a person prefers salty or sour foods 76.4% of the time. Random guessing would be correct about 50% of the time.

Q2: Sweet vs. umami flavor preference

Key findings:

• People who are more sensitive to negative emotion (e.g., cautious, higher anxiety, higher volatility) are significantly more likely to prefer sweet foods over savory foods. (p=0.0001, t=3.8765).
• People who are more likely to generally prioritize the wants and needs of other people over their own are significantly more likely to prefer sweet foods over savory foods. (p=0.0002, t=3.6776).
• People who are more likely to prioritize their own learning and growth over the wants and needs of other people are significantly more likely to prefer savory foods over sweet foods. (p=0.0037, t=-2.901). 

How accurate is the model? 

Given participants’ total trait scores, the trained model can correctly predict whether a person prefers sweet or savory foods 57% of the time. Random guessing would be correct about 50% of the time.

Q3: “Which of the five main taste types is your favorite?”

Key findings:

Whether people prioritize their own goals and well-being (self-oriented) over those of others (other-oriented) is by far the largest predictor of their favorite taste (p≈0, f=17.1272). People who reported bitter or sour tastes as their favorite were the most self-oriented, whereas people who reported sweet tastes as their favorite were the most other-oriented. This reinforces the intuitive idea that "sweeter" people may indeed have a sweet tooth!

Whether someone prioritizes others over their own learning and growth is the second largest predictor of their favorite taste (p≈0, f=11.826). Again, we find that people who reported sweet tastes as their favorite were the most likely to prioritize other people, while those who reported bitter and sour tastes as their favorite were more likely to prioritize their own growth. This replicates other research demonstrating that people who prefer bitter foods also tend to be higher in some measures of openness.

Whether people are more or less sensitive to negative emotion (e.g., extent to which a person is more cautious, anxious, volatile) is the third largest predictor of their favorite taste (p≈0, f=8.169). People who reported sweet and salty tastes as their favorites were the most likely to be more sensitive to negative emotion. This finding also replicates past research that suggests links between increased stress and salt and sugar intake.

How accurate is the model? 

Given participants’ total trait scores, the trained model can correctly predict which of the five main taste types is a participants’ favorite 44.8% of the time. Random guessing would be correct about 20% of the time.

Major takeaways

In general, this analysis suggests that the results of the Decision Style Test can be used to accurately predict information about people’s flavor preferences. There are many interesting applications of this kind of research—for instance, restaurants might consider customizing their menus to best suit the individual personalities of their customers. 

This investigation also demonstrates that there is likely some neurological mechanism that links our different types of "tastes." We can’t conclude from this analysis that personality causes particular taste preferences or vice versa; we can only conclude that the two are statistically related—and that better understanding the nature of this relationship may cast light on how this causal nature of this association. 
 

Above image courtesy of BCBST News Center.