Mircea Zloteanu
October 2020

What do you work on?

I look at how people make deci­sions in sit­u­a­tions of uncer­tain­ty and how our opin­ions are biased by avail­able infor­ma­tion. For example, our ability to detect when others are lying may be dis­rupt­ed by how we per­ceive peo­ple’s emotions.

“Are people willing to live with a serial killer if the incen­tives are just right?”

What did you do using Gorilla?

I have been using Gorilla to program exper­i­ments looking at peoples’ online behav­iour on sharing economy plat­forms. Using Gorilla, we created a sharing economy plat­form (e.g. a fake AirBnB-esque website) that offered accom­mo­da­tion to measure how people make deci­sions about hosts given dif­fer­ent trust and rep­u­ta­tion infor­ma­tion (e.g., reviews or star ratings).

Gorilla allowed us to inves­ti­gate which infor­ma­tion is pre­ferred by users on such plat­forms and the impact it has on judgement.

In the near future, we will be using Gorilla to see if users can detect the quality of online pro­files, how culture affects judge­ment, and if people are willing to live with a serial killer if the incen­tives are just right.

Has this study been published?

The initial sharing economy data has been pub­lished in Plos One.

Read our paper here.

What is the most excit­ing piece of research you’ve ever done?

A recent accom­plish­ment of mine is pub­lish­ing a project looking at how well people can detect decep­tive facial expres­sions of sur­prise. I was thrilled when it came out, because the research itself was the cul­mi­na­tion of the work I did during my PhD, uncov­er­ing why using facial expres­sions of emotion (as seen in media depic­tions of lie detec­tion) seems to not work.

What I found was two-fold. First, people are poor at detect­ing that the emotion on someone’s face is not genuine. This means that even if we learn to cat­e­go­rize facial expres­sions really well it will give us no benefit because we cannot sep­a­rate emo­tions someone feels versus those they choose to show us.

Second, we showed that emotion recog­ni­tion research com­par­ing “posed” and genuine emotion detec­tion has a severe lim­i­ta­tion, because they do not account for how emotion pro­duc­tion affects per­cep­tions. We showed that not all “posed” emo­tions are created equal, and that research needs to con­sid­er this for us to have an accu­rate picture of human emotion recognition.

“Relying on emo­tions, such as facial expres­sions, can be detri­men­tal to lie detection.”

What are the main ways people mis­un­der­stand your thesis?

Some­thing that I get asked fre­quent­ly is if I am better than other people at catch­ing lies, given that this is my field of exper­tise. The main aim of my thesis was to explain that humans are in general very bad at catch­ing lies, as they are biased towards assum­ing people tend to be honest, and because they place too much empha­sis on emo­tions as cues for decep­tion. My work actu­al­ly shows that relying on emo­tions, such as facial expres­sions, can be detri­men­tal to lie detec­tion, which is partly due to our ability to fake emo­tions really well.

How do you think online research is going to change your field?

Acad­e­mia is a very com­pet­i­tive place, with huge pres­sure being placed on researchers to publish high quality work very quickly.

Moving psy­cho­log­i­cal studies to an online domain can take some of this pres­sure away, espe­cial­ly for younger researchers whom are con­stant­ly running against the clock. Even if you decide to run a Gorilla experiment in the lab locally, having the data be auto­mat­i­cal­ly record­ed and easily export­ed to other plat­forms is a huge time-saving benefit.

For you, what is the stand-out feature in Gorilla?

Gorilla is a very ver­sa­tile tool, making its lim­i­ta­tions mostly a result of the user’s expe­ri­ence with the tool. Setting up exper­i­ments and making mod­i­fi­ca­tions to them once they are created is very easy and intuitive.

How did Gorilla make your life or research better, easier or faster?

One feature that I like is that you are not con­strained to using only the tools or tem­plates pro­vid­ed by the plat­form. If your javascript skills are good enough, you can program almost any study you want.

What advice would you give to someone start­ing out in behav­iour­al science/research?

If you truly want to have a suc­cess­ful career in acad­e­mia, you need to learn how to plan in advance. A great part of being able to stay on course and progress in your career is to ensure you have an excel­lent work ethic. You need to make sure your routine is scal­able to when you start working on mul­ti­ple projects, and are under time pres­sure. Also, make sure to engage with other aca­d­e­mics and feel free to ask them ques­tions about what they do and what tips they have; others’ expe­ri­ence and wisdom can be invalu­able in your per­son­al devel­op­ment. But also, find some­thing that can truly relax you for your downtime.

What science book have you read recent­ly that you’d rec­om­mend to others?

While this is a science-fiction book (so I may be cheat­ing on the answer), one that I come back to from time to time and would highly rec­om­mend is Ren­dezvous with Rama by Arthur C. Clarke. It is, in my opinion, an excel­lent descrip­tion of how the sci­en­tif­ic process when inves­ti­gat­ing some­thing novel should unfold. At times it can be slow paced, unevent­ful, and leaving us with more ques­tions than when we started, it can also so be surreal, illu­mi­nat­ing, and awe-inspiring.

Mircea Zloteanu
Picture showing a test glass Lie Detection, Emotions, Decision-Making, User Behaviour, Investigations
Picture showing an university graduates hat Forensic Psychology Lecturer
Picture showing a School Teesside University
Mircea Zloteanu

Ready to get started?

More Spotlight Interviews

Matthew Hunt

Matthew Hunt

[get-spotlight-info] "I used Gorilla to run a variant categorization task using a mixture of regular words, swearwords and legal non-words all ending with the morpheme ‘-ing’.” Continue Reading Matthew Hunt