I am a Ph.D. candidate in the Department of Sociology at Cornell University and a member of Michael Macy’s Social Dynamics Laboratory. Next to my studies in sociology, I pursue a minor in the Department of Computer Science with a focus on machine learning and natural language processing and I work for the OECD International Programme for Action on Climate (IPAC) as a data science consultant.

My Ph.D. committee consists of Michael Macy (Cornell Sociology and Information Science), Filiz Garip (Princeton Sociology), Felix Elwert (UW-Madison Sociology and Statistics), Eleonora Patacchini (Cornell Economics), and Lillian Lee (Cornell Computer Science). I am a research assistant of Kelly Musick (Cornell Public Policy).

In my research, I leverage and develop quantitative methods to explore social phenomena through conceptually precise and methodologically rigorous empirical research.

My substantive research aims to advance our understanding of family and network dynamics as drivers of social inequality. My methodological research aims to enhance existing methods to understand complex dependencies among observations, such as multilevel, spatial, and network embeddedness.

My research is funded by the National Science Foundation.

Quantitative methods for the social sciences (esp. causal inference with multilevel, spatial, network, and textual data) ++ social networks ++ social inequality/stratification ++ family demography

Social network analysisSocial network analysisTook several classes and workshops on network formation and network effect models and published on these topics. ++ Machine learningMachine learningTook the machine learnings sequence in the Cornell CS department (CS5780: ML for intelligent systems, CS5786: ML for data science) ++ Natural language processingNatural language processingTopic modeling, neural networks for text data (e.g., word2vec), and time-series modeling on text data ++ generalized linear and additive modelingGeneralized linear and additive modelingTook several PhD-level classes on GLM/GAM in the Cornell statistics department. Both R packages I developed are GLM/GAMs. ++ multilevel analysisMultilevel analysisTook several classes on multilevel modeling. The model underlying the R package 'rmm' I developed is a extended multilevel model ++ longitudinal & panel analysisLongitudinal and panel analysisTook several courses on longitudinal and panel analysis. The model underlying the R package 'ineqx' is a model of change over time. ++ survival analysisSurvival analysisTook several classes on survival analysis in the Utrecht biostatistics department. Experience with continuous and discrete-time survival anlysis. ++ structural equation modelingStructural equation modelingTook several classes on structural equation modeling in the Utrecht statistics department. ++ modern causal inferenceModern causal inferenceTook several classes on matching, instrumental variables, regression discontinuity, difference-in-difference designs, and the potential outcome model in the Cornell economics department. ++ Bayesian statisticsBayesian statisticsTook several classes on Bayesian statistics in the Cornell statistics and Utrecht statistics department. The model underlying the 'rmm' package I developed is estimated using Bayesian MCMC in Jags. ++ statistical computingStatistical computingExperience in statistical computing and simulation because all my methodological projects include simulations to examine the statistical properties of the developed models. ++ parallel computingParallel computingExperience in parallel computing through my methodological projects, which required extensive simulations that I sped up using cloud computing and parallel processing. ++ imputation techniquesImputation techniquesExperience with multivariate normal imputation, imputation by chained equations, and full information maximum likelihood. ++ measurement modelingMeasurement modelingTook classes on latent variable modeling (structural equation modeling) as well as classical and nonclassical measurement error. ++ survey & sampling designSurvey + sampling designTook survey data analysis, psychometrics, clinical research designs in the Utrecht Statistics department.

Programming: Highly proficient in R (>10 years) (especially tidyverse, ggplot), Python (>2years) (especially NumPy, SciPy, Pandas, Scikit-Learn), Stata (>10years), Bugs/Jags/Stan (>5years)

Experience with PHP, SQL, (X)HTML, CSS, and JavaScript.

• Presented my project “Socioeconomic segregation in friendship networks: Prevalence and determinants of same- and cross-SES friendships in US high schools.” at the 2022 meeting of the RC28 Research Committee on Social Stratification in London (April 2022), at the Measuring Belief Systems workshop at Princeton University (October 2022), at the MZES Colloquium at Mannheim University (October 2022), and at the Inequality Discussion Workshop at Cornell University (November 2022). I will present an updated version of this paper at PAA 2023 in New Orleans, Louisiana (April 2023), at Sunbelt 2023 in Portland, Oregon (June 2023), at ISA in Melbourne, Australia (June 2023), at NetSci 2023 in Vienna, Austria (July 2023), at RC28 in Ann Arbor, Michigan (August 2023), and at ASA in Philadelphia, Pennsylvania (August 2023).
• Was a visiting researcher at the Berlin Social Science Center (WZB) (June+July 2022)
• Presented my project “Treatment effects on within-group and between-group inequality. A causal decomposition approach” at the 2022 meeting of the Population of America Association in Atlanta, USA (April 2022), and at the 2022 meeting of the International Network of Analytical Sociology (INAS) in Florence, Italy (May 2022)
• Gave an introductory workshop on multilevel modeling for the Cornell Population Center (March 2022).

Cornell University
109 Tower Road
368 Uris Hall
Ithaca NY, 14853

Email fbr33 at cornell dot edu

You can schedule an appointment with me here.

Ben Rosche 0000-0001-5196-625X Ben Rosche benrosche Ben_Rosche