Consulting - Vignettes
This site shows a few examples of the projects on which we worked in the Playroom.
- International conflict management in post-1990 wars
- Poverty measures
- Interpreting multilevel varying-intercept, varying-slope models for experimental data
- Teacher/course rating
- Multistate bridge sampling methods
- Regulatory focus and regulatory fit
- Poverty in the city
- Earth Institute workshop
- ConEdison project, statistical help
- Power calculation for propensity score analysis
- Child care subsidy use in NYC
- Selecting tombs in an ancient Mexican ceremonial center
-
****
-
International conflict management in post-1990 wars
Bernd Beber (Political Science) writes: "The two undergraduate research assistants under my supervision, Matt Heiman and Michelle Forman, primarily worked on two types of tasks. First, they assisted in constructing a dataset of international conflict management in post-1990 wars. This entailed completing advanced search queries for content delivery through Factiva, reading a set of newspaper and wire reports about each conflict, and coding the information for inclusion in the dataset.
Second, for the majority of the semester, the undergraduates were introduced to geographic information systems (GIS) and learned to use ArcGIS, a powerful program for geoprocessing, spatial analysis, and cartographic design. In particular the students were taught to process vector data and raster formats, edit feature shapes and attributes, understand and work with coordinate systems, execute simple geographic inquiries, manipulate map symbology, and create basic map designs. Their applied GIS work contributed to two projects. First, they helped to vectorize enumeration area maps for a recent survey on the correlates of riot participation in Nigeria. Second, they created maps of the incidence of war and conflict management since 1990, using the dataset they helped to construct."
Poverty measures
Christine D'Onofrio (Office of Poverty Research) writes: "...it's good to hear that we're on the right track with this standard error problem. We're going to explore some differences in the standard error of our poverty rate and the official poverty rates, then try your suggestion of taking the average errors of multiple imputations and see where that leaves us."
Interpreting multilevel varying-intercept, varying-slope models for experimental data
Jean Ensminger (Cal Tech Anthropology) writes: "I spent months on the dataset after we spoke, and eventually settled on a compromise approach that would be satisfying to no purest, but was the best we could do with some of the limitations that we confronted. We have just finished drafting a short piece for Nature--about to be submitted. The book is almost finished and should be shipped off next month. Once again, many thanks for your gracious help."
Teacher/course rating
Jose Uribe (Office of Planning and Institutional Research) is looking at teaching evaluations from classes in the summer 2002 until the summer of 2008. He is working on a fixed effects model to disentangle the effects of highly-rated classes from those of highly-rated teachers.
Multistate bridge sampling methods
Michael Shirts (Univ. of Virginia, Chemical Engineering) writes: "We were able to trace down the other references, and put together a paper describing how multistate bridge sampling methods can be used effectively in chemistry and statistically physics, which was published last month in the Journal of Chemical Physics. It would not have been nearly as good a paper without the nudge you gave me in the right direction!"
Paper reference: Shirts, M.R. & Chodera, J.D. (2008). Statistically optimal analysis of samples from multiple equilibrium states. Journal of Chemical Physics, 129:124105. (link)
Regulatory focus and regulatory fit
Michel Pham (Business School) writes: "Thank you very much for meeting with me today to advise me on the analysis issues that I had regarding the following paper:
Michel Tuan Pham and Hannah Chang, "Regulatory Focus and Regulatory Fit in Consumer Search and Consideration of Alternatives." Revision invited by Journal of Consumer Research. (Research supported by a 2006-2007 Off-Term Support Grant from the Graduate School of Business, Columbia University)."
Poverty in the city
Irv Garfinkel is working with Mark Levitan at Poverty Research. They are looking at poverty in the city for Mayor Bloomberg; the ASC is helping with the statistical work (standard error construction and calculation).
Earth Institute workshop
On October 8, 2008, Columbia's Earth Institute held a workshop titled Technical Issues in Index Insurance. Index-based financial risk-transfer mechanisms are being tested in the context of development and climate-change adaptation. These instruments show the capacity to dramatically reduce conditions of chronic underdevelopment by both enabling investment and reducing shocks in agricultural livelihoods. However, important issues have been raised, such as the potential for scaling up and the impact of climate change on index insurance. The workshop brought together the research and implementation communities to discuss how state of the art research and technology can best serve development, today.
ConEdison project, statistical help
This project involves decision making by Con Edison, using causal inference, statistical analysis, machine learning, and survival analysis. The leader of the project is Dave Waltz in the Center for Computational Learning Systems; other people on this project are Andrew Gelman, Zhiliang Ying, Samantha Cook, and several Ph.D. students.
Power calculation for propensity score analysis
Ilan Meyer, Associate Professor at the School of Public Health, is working on a grant which needs statistical help from Andrew Gelman on power calculations for propensity score analysis.
Child care subsidy use in NYC
Margo Gardner, Jeanne Brooks-Gunn, and Anna Johnson are part of team working on this project. Andrew Gelman is working on this as a consultant, giving statistical guidance and analysis. The project is contacting low-income working families in NYC to ask them about their use of voucher and non-voucher subsidized child care.
Selecting tombs in an ancient Mexican ceremonial center
Ellen Hoobler is working on her thesis, involving tombs in an ancient Mexican ceremonial center. She needed assistance with the statistical aspects of the selection process for the tombs on which her thesis focused.