The Center on Poverty and Social Policy at the Columbia University School of Social Work, the Columbia Population Research Center, and the Institute for Social and Economic Research and Policy are seeking a postdoctoral scholar with a PhD in statistics, economics, political science, public policy, demography, psychology, social work, sociology, or a related discipline, to lead the development of survey weights for the New York City Longitudinal Study of Well-being and associated datasets. The scholar will also help estimate empirical models and contribute to reports of research findings. The postdoc will work closely with professors Irwin Garfinkel (PI) and Andrew Gelman (co-Investigator), project director Christopher Wimer, and survey director Kathryn Neckerman. Additionally, the postdoc will collaborate with Andrew Gelman on designing methods for model-based analysis of complex surveys.
The ideal candidate will have a strong background in statistical modeling and an interest in working on methods for survey analysis — including both design- and model-based approaches — and have strong programming skills in R and willingness to use Stata. Previous experience with the statistical modeling language Stan is a plus but is not required.
Please submit the following materials, which will be reviewed on a rolling basis, to [email protected]:
CV or resume
Sample paper or publication
A list of three references who can speak to the candidate’s credentials
Applications will be reviewed on a rolling basis. The priority deadline is January 31, 2020, but we will continue interviewing until we find the best-matched candidate.
Each application will be evaluated on the following criteria:
The degree of fit of research interests
How the postdoctoral training will help to further the postdoctoral scholar’s research trajectory
Demonstrated record of work with quantitative data and statistical programming
Demonstrated quality of writing.
When we say “including both design- and model-based approaches” of survey analysis, what we really mean is that design- and model-based approaches are the same thing.
Design-based approaches all assume models, and model-based approaches should account for design (as discussed here).
What we’re working on is improved design-based adjustments that make use of better models. Or, to put it another way, better model-based adjustments that account for design. And also better design that anticipates the adjustments that will be done.