(Asian)

Bias in

Dating Apps

Emily Zou & Nandhini Gounder

FATE

Fall 2020

INTRODUCTION

We thought the topic of bias in dating app algorithms would be interesting to take a look at after Jason suggested it to us. There is also not much research into the topic, save for a few high profile studies. The topic of Asian bias in dating apps has not been specifically investigated either.

Background

Research

Results

Discussion

Topics

01

02

03

04

Knowledge going into the project, news and articles

Research methodology and reasoning

What we found through our survey and interviews

Analysis of results and insights

BackGround

01

just some casual google searches, everyone does it right?

Readings

• Race filters can aggravate racial bias on platforms
• Black women and Asian men were likely to be rated lower than other ethnic groups - 2014 OKCupid study
• People with no ethnicity preference still received a majority of matches with people from their own ethnicity on Coffee Meets Bagel - 2016 Buzzfeed report
• Bias in using historical data to extrapolate matches for users - becomes a self-fulfilling prophecy -> more likely to date a specific race if you are predominantly shown matches in that race, does not account for factors that lead to past matches, does not provide space for new/emerging factors that could lead to matches - 2018 Cornell study

REsearch

02

time to get the deets

Research Questions

• Who is being shown by these dating algorithms?
• In particular, what kinds of people are Asians getting?
• What are people’s perceptions of bias in dating app algorithms?

Methodology

• Created a survey on Qualtrics
• $42 on Amazon Mechanical Turk for 50 responses
• Posted on Facebook groups
• Conducted semi-structured interviews to get more in-depth insight

Survey

• Get as much information as possible within a limited time frame
• Asked questions regarding:
• General dating app preferences
• Experience with bias
• Asked participants to swipe 10 times on their preferred dating app and record each profiles race/ethnicity to the best of their knowledge

Semi-Structured Interviews

• Difficult to get more nuanced views through mass surveys
• Data from survey is more likely to be lower quality in terms of responses (people are trying to fill out the survey as fast as possible)
• Asked 3 participants to talk more in-depth about their dating app preferences and experiences specifically in terms of bias
• Explained the algorithms behind popular dating apps and asked for their opinion now that they knew this information

REsults

03

so is there bias? in short, yes.

Survey REsults

• 197 total responses
• Cleaned up data - eliminated responses that weren’t completely finished
• 130 complete responses: 50 from MTurk and 80 from Facebook
• Mostly men; mostly in the 16-25 and 26-35 age groups
• Mostly positive view of dating apps

Match.com

Tinder

Bumble

3rd

1st

2nd

Plenty of Fish

4TH

Most Used Dating Apps

Who is being shown?

Who Does the Algorithm show to Asian people?

A STudy in 2013 - Dating prefERences

Our study in 2020 - Who Gets SHown

Women

Men

Non-Binary People

WhAt are people’s perception?

Interview REsults

• Initially only going to interview 2 people; ended up getting a 3rd person (the more the merrier amiright?)
• Demographic info:
• East Asian; 25; Female
• Turkish; 24; Female (white-passing, but does not prefer to identity as white)
• East Asian; 23; Female

Monster Match

Dating apps use your own preferences and the preferences of the people around you (similar to how Netflix recommends movies) to show you specific people - i.e. less popular people are not shown as much. How do you feel about this?

​

General Sentiment:

I feel bad for the people whose profiles aren’t shown as much…

Aren’t dating apps supposed to help you find people?

​

Overall Thoughts on Bias

• Participants were aware of preferences for Asian women due to certain problematic stereotypes
• Participant 3 was aware of bias against black women on dating apps - “I’ve read that black women are the least desired group on dating apps”
• Participant 3 also mentioned men, and Asian men specifically having a harder time finding matches on dating apps due to the gender skew and problematic stereotypes
• All participants mentioned that dating apps could be improved by making it more fair for all groups

​

Discussion

04

some general thoughts and insights

People are more okay with the idea of bias in dating apps because they want the apps to accurately reflect their own preferences in the search process so they get matches more reflective of those preferences.

BiAs

The UI of the app often creates, informs, and promotes the type of dating attitudes prevalent on the app.

Culture

People want a balance between specificity to weed out people that will waste their time and having a large enough pool people so they have options.

Niches

Insights

Interesting Tidbits

• Tinder creates a ranking or ELO for you based on the number of swipes you get - this is an attractiveness score. As your ELO increases, your profile is shown to people with higher ELOs
• Tinder Gold algorithm and queuing system - allows you to change your location and move to the top of the queue in that location.
• The importance of liking the same foods for a long term relationship was a common

theme across all interviews.

• Men in Missouri like to pose with fish.

THANKS!

Does anyone have any questions?

​

CREDITS: This presentation template was created by Slidesgo, including icons by Flaticon, and infographics & images by Freepik.

​

​