Research

In the summer of 2020, I worked with Professor Ed Novak in mobile security research. Specifically, our research project is called Stigma: Portable Dynamic Information Flow Tracking for Android. Decoding those jargon-y words, Stigma allows you to track how information flows through your smartphone apps and derive insights about the security of your apps.

The summer was when the Covid pandemic had just hit and everyone had to rapidly adjust to a new way of life. I had to also find a place to live and work with Professor Novak remotely. Thankfully, I was able to do this and the experience I had was truly illuminating and fulfilling. Our work is on-going and I hope it can be used to analyze the behaviour of apps and understand how secure they are.

This past summer I worked on the research project “Coding Media Coverage of Presidential Campaigns” with Professor Stephen Medvic as a Hackman research assistant. The question behind the project was “How effective are presidential campaigns in the US?” Professor Medvic had conducted weekly surveys of election analysts during the 2008, 2012 and 2016 campaigns asking them to select the candidate who had run a better campaign in the previous week. Weekly campaign performance should influence a candidate’s standing in the following week’s public opinion polls; however, we had to control for media coverage in order to verify that positive or negative coverage of the candidates was not the real driver of poll results. My responsibility during the project was to read through all front-page articles from the New York Times for each day during the campaigns in all three years and code the tone of coverage for each candidate.

This project looks at how children with Neurofibromatosis type 1 (NF1) fit in with cognitive load theory. Cognitive load theory was developed by John Sweller in 1988, and is a model for how individuals learn new information. When learning something new, information is first stored in working memory, where it is rehearsed and practiced (like a new math problem, for example), and then is moved to long term memory. Cognitive load is broken down into three categories; intrinsic, extraneous, and germane. Intrinsic load refers to how complex the task is, extraneous load refers to how the task is presented, and germane load refers to schemas in which information is broken down into to store in long term memory. In general, you want to simplify intrinsic load, reduce extraneous load, and maximise germane load. By doing so, you prevent cognitive overload. Cognitive overload is when too much information is received, and the new information does not make it to long term memory. For normally functioning children, they can learn strategies to simplify, reduce, and maximize these three elements to learn new information, but children with NF1 are not able to do so; they require assistance in making plans that break down information, and help them learn it.

This project uses a test called the TEC, which is a computerized task that allows researchers to understand where cognitive overload occurs, and why. It consists of three, increasingly complex tasks, and produces a lot of data to analyze. The goal of this project is to see where cognitive overload occurs, why it occurs, and specifically, where children with NF1 fall compared to normative children. At the end of the project, we hope to submit a paper to a few psychology journals, and a neuropsychology conference, that will occur next year.

I worked with Health Services Trust, which is a South African research-oriented non-profit based in Durban that’s focused on the development of health systems. I worked with public health data related to one of their annual publications, the District Health Barometer. The specific statistic I worked with was the Couple Year Protection Rate, or CYPR. CYPR is a composite measure of how often different types of contraceptives are used and represents how many years of protection the average couple has. It can also be used to see what contraceptive methods are popular in a given region and where contraceptive distribution/education can be improved. My main task was to produce and analyze the 2019/20 yearly CYPR data and look into any trends I found.