Computational modelling: I model human language learning as a discriminative process of uncertainty reduction, using the Rescorla-Wagner learning rule. Here is an example simulation in which I implemented this in Python. The model was used to generate predictions about human learning, which I tested in a series of experiments (see below).
Online experiment for adults: This is the program I wrote in javascript using Gorilla, a platform for on-line behavioural data collection. I use this and variations of this experiment to collect a lot of data with adult learners for my PhD project. Please contact me if you would like a copy of this program.
Computer game for children: In one of the projects I am involved with, we test predictions about discriminative learning in the context of second language learning. For that purpose, I developed a computerized training game for teaching 7-year-olds Japanese spatial constructions. This was done in PsychoPy, a Python library incorporating pygame. Watch this space for code!
Pre-registration of hypotheses: It is good scientific practice to plan your data analysis before you start the actual data collection. Not only does this prevent practices such as p-hacking and data-tweaking, but it is also a valuable opportunity to think further about your hypotheses and the types of statistical methods that you can use to test those hypotheses. Here you can read analyses plans for some of my experiments.
Statistical analysis: Here you can find the code used to analyse data using mixed effects logistic models and Bayes Factors for this publication.
My colleagues at the Language Learning Lab do research with children in primary schools. To speed up the process of school recruitment, I built a webscraper to automatically extract primary school contact details from council website in the Greater London area. Here is an example for the Borough of Haringey.
A PsychoPy program for soundfile transcription. The program plays soundfiles, collects transcriber’s transcription, and logs all the relevant information in a csv file. I wrote this program for the colleagues in my lab who used production tests to test participants’ learning at the end of a language training session, and needed a way to transcribe participants utterances efficiently.
As a graduate researcher at UCL, I delivered several invited talks, workshops and tutorials on a variety of topics in research methods.
Bayes Factors with Mixed-Effects Models in R, presented at the LiLaC Lab Meeting, UCL, November 2018
Basic data handling and using R and R studio. Par 1 of a three-part R-workshop co-organised with Dr Lee De Witt and Dr Catriona Silvey for UCL Chandler House staff and students, November 2018
Quick introduction to online experiments with Gorilla, presented at the Early Career Researchers Meetings at UCL Department of Language and Cognition, February 2018
© 2019 Maša Vujović. All rights reserved.