CSSTP CALL FOR NOMINATIONSYEAR 1 COHORT– 2020/2021CSSTP PROGRAM DESCRIPTION
Overview
The Computational Social Science Training Program (CSSTP) is a new two-year multi-disciplinary training program in advanced data analytics for predoctoral students in the social and behavioral sciences. CSSTP aims to prepare social science researchers to tackle the complex health problems prioritized by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), including maternal and child health, adolescent health, pubertal timing, mental health, health disparities, and the social determinants of health.
CSSTP combines Berkeley’s long-standing strength in quantitative social and behavioral science with its nationally-recognized campus programs in data science education, practice, and research. It serves five trainees per year over five years. The training faculty includes 21 social scientists who have exemplary records of developing and applying novel statistical methods to health-related social/behavioral science problems, as well as 13 data scientists who are leading figures in the foundations of mathematics, statistics/biostatistics, and computer science.
Eligibility
Second- and third-year predoctoral students who
- are enrolled in your PhD degree programs: Sociology, Demography, Social Epidemiology, Public Health Policy, Social Welfare, and Public Policy
- have completed the first-year course requirements in their home departments
- are US citizens or permanent residents
- are eligible for the two-year CSSTP.
Program Goals
CSSTP emphasizes a team science approach to problem solving and prepares students to apply novel methodologies and data analytic techniques in behavioral and social sciences research. Trainees can expect to acquire the following core competencies during the two years of the program:
Methods and tools for causal inference with observational data, particularly longitudinal data with time-varying confounding, adaptive interventions, and mediation analysis with longitudinal data;
Unsupervised and supervised machine learning algorithms and their application to health data, particularly the relationship between conventional regression models and supervised machine learning, and use of non-standard loss functions to harness supervised machine learning for problems of particular relevance in the social sciences (such as identification and response to effect heterogeneity);
Principles, methods, and applications of text analysis and natural language processing;
Expertise in responsible conduct of research, including tools for reproducibility in research and best practices for transparency and open science;[DM2]
Proficiency in tools and methods for research with high volume, high intensity, and non-rectangular data, including data manipulation and transformation, cloud computing, and data security;
Effective written and oral communication in preparation for article writing, grant writing, presentation of research results, and teaching.
The program accommodates the requirements of each of the constituent PhD programs while providing sufficient flexibility to explore specific interests through an individualized training plan. This built-in flexibility and careful sequencing of required elements ensures that trainees’ time to degree is not delayed.
This diagram illustrates the program core competencies and program design:
Program Design
Primary components of CSSTP include
- A new year-long course in computational social science that prepares trainees to employ advanced data analytic methods;
- Two data science elective courses selected from existing courses to develop deep specialized knowledge and skills;
- Data science research internships in UC Berkeley faculty labs and/or industrial labs; [DM3]
- A weekly Computational Social Science Workshop (CSS Workshop) where faculty and trainees give/receive peer mentoring, focus on professional development, discuss new research articles, are supported in paper writing and article submission, and receive feedback on ongoing research;
- Participation in regular professional development activities, including the Computational Social Science Annual Meeting (CSS Annual Meeting) to present preliminary work and receive feedback, and data science conferences to share research and begin building a professional network;
- Responsible Conduct of Research training, involving both principles and tools and emphasizing reproducibility in research, at multiple points in the training program and integrated into other primary components, including the CSS core course, CSS Workshop, and internship.
- Joint mentorship by both social science and data science training faculty;
- A research focus on intensive or voluminous longitudinal data and data from high-density, large sample or population level agency databases Trainees who successfully complete components 1-6 will receive a Certificate in Computational Social Science that documents trainee competencies in advanced data analytics and distinguishes them from other social science PhDs and makes them more competitive for postdoctoral and other positions after graduation.
Year 2 of grad studies; Program Year 1:
Close to the beginning of fall instruction, trainees attend a program orientation that covers program requirements and structure, advice on connecting with data science faculty, requirements for responsible conduct of research training, review of individual development plans, and an introduction to key program leadership and training faculty. Initially, each trainee is matched with one social science mentor and one data science mentor based on their research interests and developmental goals, although trainees are allowed if necessary to change mentors as the program progresses. Students who may need extra preparation and support for success in CSSTP aree counseled by the co-Directors on necessary BIDS or D-Lab training. During Program Year 1, trainees take the two semester CSS Core Course, attend the weekly CSS Workshop, RCR training, and the CSS Annual Meeting, and the BSSR Data Analytics T32 cross-site grantee meeting in Washington DC. Additionally, trainees in some departments are required to complete a master’s paper or equivalent in year 2 of grad studies.
Year 3 of grad studies; Program Year 2.
Trainees engage in a two-semester internship in a data science lab/research group on campus or externally in industry, and complete two elective courses in data science. They also complete their home department’s Qualifying Exam (a requirement for all Berkeley PhD students) and continue their involvement in the CSS Workshop and CSS Annual Meeting. A central goal of this year in CSSTP will be ensuring all trainees submit at least one paper for publication by the end of the year.
This diagram further illustrates the program design:
Program Directors and Training Faculty
CSSTP is co-directed by David Harding, David Mongeau, and Maya Petersen. The co-Directors have varied and highly complementary expertise in biomedical training, curriculum development, project management, and diversity/inclusion.
- David J. Harding is Professor of Sociology, Faculty Director of the Social Sciences D-Lab, and a BIDS Senior Fellow.
- David P. Mongeau is Executive Director of the Berkeley Institute for Data Science (BIDS).
- Maya Petersen is Associate Professor of Biostatistics/Public Health and co-Chair of the UC Berkeley Graduate Program in Biostatistics.
Training Faculty from the Social Sciences are :
Faculty | Department |
---|---|
Adrian Aguilera | Social Welfare |
Jennifer Ahearn | Epidemiology |
Henry Brady | Public Policy |
Julian Chow | Social Welfare |
William Dow | Health Policy |
Dennis Feehan | Demography |
Josh Goldstein | Demography |
David Harding | Sociology |
Heather Haveman | Sociology |
Hilary Hoynes | Public Policy |
Sol Hsiang | Public Policy |
Alan Hubbard | Epidemiology |
Rucker Johnson | Public Policy |
Jenna Johnson-Hanks | Sociology |
Amy Lerman | Public Policy |
Mahasin Mujahid | Epidemiology |
Ziad Obermeyer | Health Policy |
Emily Ozer | Epidemiology & Health Policy |
Jessie Rothstein | Public Policy |
Daniel Schneider | Sociology |
Jennifer Skeem | Social Welfare |
Susan Stone | Social Welfare |
Training faculty from Data Science are:
Faculty | Department |
---|---|
Joshua Blumenstock | School of Information |
Karen Chapple | City & Regional Planning |
Peng Ding | Statistics |
Anca Dragan | Electrical Engineering & Computer Sciences |
Sandrine Dudoit | Statistics |
Laurent El Ghaoui | Electrical Engineering & Computer Sciences |
Avi Feller | Statistics, Public Policy |
Michael Jordan | Electrical Engineering & Computer Sciences |
Michael Mahoney | Statistics |
Fernando Perez | Statistics |
Maya Petersen | Biostatistics |
Sam Pimentel | Statistics |
Phillip Stark | Statistics |
Mark van der Laan | Biostatistics |
Nomination Requirements
To nominate a student for CSSTP, please submit:
- a one-page nomination letter that addresses the following:
- student’s ability in quantitative social science as demonstrated through coursework, research experience, and/or teaching;
- student’s successful completion of first-year requirements in your home PhD program
- support of one or more home PhD program faculty
- a one-page statement from the student addressing the following:
- enthusiasm for participation in a multidisciplinary intellectual environment and computational social science research related to health or the social determinants of health, broadly defined.
- alignment between career goals and program goals, particularly a strong interest in a career in health research after graduation
- a copy of the student’s undergraduate and graduate transcripts (unofficial copies are accepted)
- the student’s current e-mail address
- in a single email to css-t32@berkeley.edu by June 21, 2020. This deadline may be extended until all positions are filled. CSSTP appointments officially begin on August 19, 2020.