bootstrap website templates

Research for impact

Science is a team sport. My role in research teams is to figure out how to take an existing line of research and turn it into an intervention. As such, my research domains are very diverse. What holds these projects together is the hope of creating knowledge and tools that can solve real problems. For links to publications, see my CV  

COVID States

Since April 2020, the Covid States project, led by faculty at Northeastern, Harvard, and Rutgers has been fielding surveys representative of all 50 states montoring the pandemic as well as government and public's reaction to it. My role on the project is validating the survey's representativeness and establishing a new standard for survey reporting. The goal is to change how scientists and journalists evaluate the extent to which a survey is representative or not.

The Disinformation Monitor

Disinformation is not a new phenomenon, but its prevalence and weaponization represents a threat to modern communication networks and media systems. The risks posed by what the Oxford Internet Institute has called the industrialization of disinformation are borne by everyone. Today, only large, well-resourced, and scientifically staffed companies and government agencies have the capacity to combat disinformation. Our goal is to build a tool that gives everyone the ability to identify and monitor disinformation on their networks, an anti-virus software for disinformation.

Building Healthy Relationships

Despite women’s increasing participation in the paid labor force and the growing outsourcing of household chores, American women consistently do more housework than men. This inequality restricts women’s ability to work for pay and lowers their earnings, compromises women’s sleep and leisure time, and lowers women’s perceptions of relationship quality. While many scholars study inequality, none that we know of have sought to fix it. In this project, we look to identify interventions that increase equality within couples by increasing couples’ belief that they can share the work equally and improve the actual balance of housework. Working with newly cohabiting couples, we develop prototypes for interventions and field them with couples to see how well they work. 

AI for Effective Practice

Much human knowledge is “knowing how” rather than “knowing that.” This behavioral knowledge is implicit, mastered through practice, trial and error, and repetition. However, in many domains of expertise, there is little formal or structured behavioral education and, where it exists is usually in the form of coaching and simulation. In this line of research, we seek to build AI which can model expert practice and provide feedback that measurably improves the learner's behavior. Our first application area is teaching, where we use rated video recordings of teachers to build predictive models and will work with schools to evaluate the effectiveness of the Artificially Intelligent system. 

© Copyright 2024 Jason Radford. All Rights Reserved.