Research

In my PhD on Learning Analytics, I'm focussing on understanding how students using Education Technology get influenced by Social Comparison. Past research has found that Social Comparison, that is, the natural and automatic human behavior of comparing ourselves with those around us is driven by the motives of self-appraisal, self-enhancement, and self-improvement. Social Comparison is a strong motivator.

Leaderboards, badges, charts - they are everywhere. They invoke Social Comparison. But... but they do not always work!

Students with different traits, goals and expectations respond different to similar social stimuli. It is possible that what motivates you may demotivate your closest friend. This motivates my search for personalizing these systems so that everyone receives a fair feedback that works.

Bright living room with modern inventory

If you are a student at Utrecht University and you would like to work on similar problems for your thesis, feel free to reach out to me. We can discuss ideas and expand it to a project that impacts the effectiveness of the process of learning. Masters students would be registered with my supervisor Dr. Sosnovsky as the first supervisor and I will be the daily supervisor. Bachelors OZP students will be supervised directly by me. If you are not affiliated with Utrecht University, feel free to reach out and we can discuss your project.

Comparison (De)motivates Students

You must have heard about Mastery goals and Performance goals. People who are mastery oriented focus on getting into depth of what they are learning and doing. They want to master the subject. Performance oriented individuals are motivated by a more tangible goals like getting a good grade or being more successful than their peers.

Though it is believed that comparing with the classmates motivates Performance oriented students, we found that a large number of performance oriented students in a beginner programming course looked at the progress of their peers, and got demotivated when they noticed that the average progress of the class is already low.

This should make us reflect on how we're designing the leaderboards and other feedback tools that compare students with each other.

Can Computers Understand Hi-nglish?

Back in 2014, I was working on a freelance project to extract Social Media users' perceptions about my client's products. I found that a large amount of data on Facebook groups and pages was not English, but Hindi words written in Roman Script. So far, text classification had been language specific. Traditional methods like converting each hinglish word to it's hindi equivalent was not doable because of variants of spellings.

I collected a dataset from popular Facebook groups and prepared a dataset for sentiment classification. I collaborated with colleagues at IIIT Hyderabad and we created an architecture that combined CNN and LSTM to capture the variations while preserving the sequence of what is being said. This proved to be a significant improvement than any workarounds of the traditional methods, and promoted further research in deep neural architectures.

Read more on ACL:
https://aclanthology.org/C16-1234.pdf

How do Students Compare?

Comparing with the peers has been mentioned as a strong motivating force in marketing, social media, health, and of course, learning. Watching your friends' progress leads to you want to study more. In this mixed-methods study with our colleagues and students, we found that students' personal attributes, intentions and goals affect how do they perceive different types of leaderboards.

worm's-eye view photography of concrete building