This portfolio is private

Enter the password to continue. If you don't have one, get in touch and I'll send it over.


← All work

Case study 03 · Published research

What influences fairness behaviours in humans?

Prior research found people made fairer decisions in an economic game when making decisions via a smiling voice agent than directly with another person. I ran the study that tested whether this effect holds for multimodal agents, like social robots, from research question to publication.

My role
Sole researcher, end-to-end
Built with
Unity, C#, HTML, CSS, JavaScript
Participants
80+ online respondents
Result
Published paper on Journal on Multimodal User Interfaces

Research background

Previous research suggested that people made fairer decisions in an economic game when they made decisions via a smiling voice agent than when interacting directly with another person. It was unclear whether this effect would extend to multimodal agents, with both a voice and face, driving the research question.

My role

I executed the project end-to-end: from identifying the research opportunity through an extensive literature review, to planning the timeline, designing the methodology, and publishing the findings.

  • Designed the research methodology and approach
  • Developed a browser-based economic game to measure user behaviour, using Unity, C#, HTML, CSS and JavaScript
  • Recruited 80+ participants — writing screening criteria, advertising the study, and handling inquiries
  • Analysed the data in SPSS and wrote up the final report

"Technical fluency expanded what I was able to investigate, prototype and validate as a UX researcher."

Building the testing environment

One of the most valuable parts of the project was learning to build the game environment needed to answer the research question, rather than treating development as separate from research and design. That meant translating an experimental methodology into a working interactive system — from designing the interaction flow to implementing game logic and capturing participant behaviour.

While it is far from a sophisticated game, it gave me hands-on experience translating research requirements into technical requirements, picking up unfamiliar programming concepts independently, building interactive experiences in Unity, and debugging a working prototype through iteration. Try out the game below!


Test ID: 300000

Outcome

Participants were fairer when making decisions via a multimodal agent than when facing their counterparts directly — a finding that speaks to how humans perceive and respond to agent emotion. The project achieved a first-class result and was published in the Journal on Multimodal User Interfaces.