GEOG 4023/5023 Spring 2008
Research Project
Students are required to perform a statistical analysis, present the results to the class at the end of the semester, and write a final paper. The goal for this research project is to provide you with hands-on analysis experience and the opportunity to critically evaluate the strengths and weaknesses of a particular statistical method or model.
There are 2 options for the research project:
Option 1: Replicate a study
Select a recently published research article that has a spatial/geographic component to the problem and whose data are available either publicly or by contacting the author. For this project, you’ll need to both replicate a key table (and figure if appropriate) in the article, and extend the analysis in some way. Your Method Proposal should suggest 3 possible improvements to the analysis. You must have the data necessary for replication in a usable geographic form by the proposal due date. See the replication data sources for a list of journals and websites that provide replication data.
Option 2: Original research
Perform a statistical analysis with real data and meaningful results. This option is for those students who have a specific research question in mind and are prepared to apply quantitative methods to that research question. You must have all of your data in a usable geographic form by the proposal due date.
Step I: Topic Proposal. Due Thur, Feb 21.
You must submit a 1-2 page proposal describing your research/replication problem. Include the justification (why study this topic), hypotheses (if appropriate), and data (do you already empirically derived spatial data or will you generate synthetic data?). Both proposals will be evaluated on a stop-light system, where you must earn a green-light before proceeding. Yellow- & red-lights will require submission of a modified or completely new proposal, respectively. [Note: if you are having trouble with this assignment, please come see me!]
Step II: Method Proposal. Due Thur, Mar 13.
Indentify the statistical method(s) or model(s) most appropriate for your research problem. Multivariate regression and the spatial variations we discuss in class are one option. Explain what the method is designed to do and what questions it can help to answer for you. You may find it helpful to read ahead in the textbooks and the Reference section of the course website to help with your proposal. Also indicate which software package(s) you will use. I recommend that you use one or more of the course software packages for your analysis, but you are not limited to these software tools if you have your own preference. At this stage, you must have your spatial dataset in a usable form (e.g. shapefile). As part of this second proposal, indicate the status of your data.
Step III. Conduct the Analysis.
Some lab time will be available to assist you in your analysis. As you begin to see results, be sure to consider competing alternatives (possibly, simpler ones!) to your chosen method of analysis (e.g. how does your spatial regression model perform relative to OLS regression?).
Step IV: Presentation. Tue, Apr 29 & Thur, May 1.
The presentation should be about 10-15 minutes (depending on the number of students), including time for questions. Please email me your presentation (e.g. Powerpoint file) one hour before class so that I can load all presentations on my laptop.
Step V. Final paper. Due in my mailbox Fri, May 2, 5 pm.
Review and rewrite the material covered in the first two steps of the project. Submit a 12-15 page term paper detailing your analysis and results (double spaced, excluding graphics; double-sided printing is encouraged). The research project is worth 20% of your total course grade. Presentation quality/clarity will be factored in to your overall project grade.
Example paper outline.
I. Introduction & Background
Define the problem
Description of hypotheses
Existing theory that informed your hypotheses
Summary of related work published on this topic
II. Methods
Review/critique of relevant statistical methods
Description of the method(s)
III. Data Description
Sample
Population
Unit of analysis
Dependent and independent variables
IV. Results of Data Analysis
Descriptive/exploratory results
Modeling results
(for Option 1: the replication analysis & your improvements)
V. Conclusions
Do the results support your hypothesis?
Shortcomings of analysis
Possibilities for further work
