Curriculum Vitae                       Contact Information:                 323 Uris Hall             Cornell University             Ithaca, New York 14853-7601             mws24(at)cornell(dot)edu             (607) 254-8922

Areas of Interest:
Sociology of Education, Inequality, Quantitative Methodology, Sampling, Social Networks

Projects:
I am currently working on a dissertation examining how cultural and family norms impact inequality of educational achievement. I will use the National Longitudinal Survey of Youth 1997 Geocode Supplement to examine such factors as fertility, age of marriage, and student willingness to move away from home for college.  I am working with Steve Morgan, Jenny Todd, and Theodore Leenman on a project examining curriculum tracking and educational outcomes.  Additionally, I am working on a project examining the effectiveness of different financial aid configurations on college completion rates for resource-constrained students.

I have been a research assistant on the 3-City Unregulated Work Study for the past three years.  This project recently released the first data report using the survey data, which has been extensively covered in the media.  I will continue to work on papers using this data with various members of the project team, and I am available for methodological and statistical consultation on projects employing Respondent-Driven Sampling.

Works in Progress:

        "Poverty, Financial Aid, and College Completion: The Best Laid Plans of Mice and Men."

       "Analysis of Respondent-Driven Sampling Data: A Fresh Look at the Bootstrapped Variance Estimator.”

        “Causal Effect Estimation with Respondent-Driven Sampling Data.”

Masters Thesis: "Regression Modeling of Respondent-Driven Sampling Data"
Respondent-Driven Sampling (RDS) is a snowball-type sampling method used to survey hidden populations.  To date, analyses of RDS data have consisted of estimating population proportions and their variance because of the special complexities RDS data pose for regression analysis.  My master's thesis discusses those complications, focusing on the role of homophily (differential affiliation) in the recruitment process and respondent clustering at multiple potential levels of aggregation.  It proposes two techniques for confronting these problems: entering recruiter characteristics directly into recruit-level regression models and estimating fixed- or random-effects models at the levels where significant clustering is observed.  An empirical example demonstrates the modeling process (in the appendix), and a six-step procedure for regression modeling of RDS data is presented. (Appendix here.)