The Department of Statistics offers a Summer Internship in Applied Statistics to undergraduate students who will be enrolled at Columbia College, Columbia School of General Studies, Barnard College, or Columbia Engineering in fall 2011.

Interns take part in the Department consulting service and participate in mentored research with Department faculty.  The internship takes place from May 23d through July 1st.  Interns receive student housing and a stipend of $3,000.

Review of applications will begin April 2nd.  Applicants should send a transcript with a cover letter to:

Ms. Dood Kalicharan
Department of Statistics, Room 1005
School of Social Work Building
1255 Amsterdam Avenue

In the cover letter, please indicate your interest in one of the project areas described below.  Preference is given to students with advanced course work in statistics, and to repeat applicants.

1.  A project examines the impact of different licensing options for scientific publication - all of articles, code, data. The intern would help design the experiment and build and manage datasets. We would use then these data to model scientific re-use of published material, under the various licensing options, to understand barrier to effective knowledge transmission.

2.  Demography and geography: In “Red State, Blue State, Rich State, Poor State,” we looked at how the patterns of income and voting vary by state; see here:

http://www.stat.columbia.edu/~gelman/research/published/rb_qjps.pdf

Recently we’ve been fitting more complicated models with nonlinear patterns that can vary by ethnicity as well as income, within each state: http://www.stat.columbia.edu/~gelman/research/published/spp_revised.pdf

Our new idea is to apply these methods to questions other than voting, in particular public health issues such as smoking and obesity.  For example, this recent map shows obesity by state:

http://healthyamericans.org/reports/obesity2010/ It would be fascinating to see how obesity varies by ethnicity and income within each state.  And if we can get the data, we can just push the button and do the analysis.  The idea would be to do this on a lot of different health behaviors and see what we find.

3.  Social and political polarization: To what extent to Democrats mostly know Democrats and Republicans mostly know Republicans.  How much are social networks politically segregated, compared to segregation by ethnicity, age, and income? Fragmentary evidence is available on this from a number of different surveys, including modules of the General Social Survey, the National Election Study asking respondents to supply information about their closest friends, and a recent module of the General Social Survey that we constructed, asking a range of questions about acquaintance and trust networks.  The challenge is to focus on the question of political polarization of social networks and attack it using a variety of data sources.

4.  Red County, Blue County in California:: In “Red State, Blue State, Rich State, Poor State,” we looked at how the patterns of income and voting vary by state; see here:

http://www.stat.columbia.edu/~gelman/research/published/rb_qjps.pdf

5. Development and application of statistical methods in studies of the genetic epidemiology of neurological, cardiovascular, and pulmonary disease.   Possibilities for interns include data analysis, programming, and methodological research.   This option is for students considering a career in biostatistics, or looking for experience in bio-medical applications.

6. Study of the impact of patents on code (and data) transparency in scientific publishing.   We'd start with the Columbia community and move to others as we could.