Draft syllabus for NYU graduate course on designing surveys & field experiments

As mentioned in a previous post (link), I will teach a graduate course this Spring at NYU called “Quantitative Field Methods.” The course description is as follows:

POL-GA 3200 Quantitative Field Methods (4 points) Instructor: Cyrus Samii (GSAS/Politics), Spring 2012, Thu 4-6
This is a graduate course on statistical methods for designing quantitative social science field research, including sample surveys, field experiments, and observational (quasi-experimental) studies. The purpose of this course is to train graduate students in the social sciences to design rigorous quantitative micro-level fieldwork for their research. The learning goals are (i) to understand why some sampling, experimental, or measurement techniques are to be preferred over others, (ii) to be able to analyze design alternatives and implement sampling, treatment assignment, and measurement algorithms in the R statistical computing environment, and (iii) to develop an ability to take meaningful social science questions and translate them into hypotheses and research designs that can address the questions in a compelling manner.


A working draft of the syllabus is here: quant field methods 120103. I welcome comments or suggestions.

For NYU students or NYC-area students interested in the course:

  • First, I recently noticed that in the registration system this class had been mislabeled (something about “design of institutions”). As far as I know, this has been corrected. In any case, you can be sure that POL-GA 3200 refers to the class described in this post and not the institutions class.
  • Second, the course is open to PhD students as well as master’s students, subject to availability of slots, from across the social sciences. (See syllabus for details.) However, the prerequisites are at least a year of graduate level social science statistics training (or equivalent, subject to my assessment), as the presentation will be technical and the assignments will involve fairly intensive programming in R. No auditing will be allowed.
Share