Skip Navigation

Carleton College

  • Home
  • Academics
  • Campus Life
  • Prospective Students
  • Alumni
  • Faculty & Staff
  • Students
  • Families

Building an Electoral Dataset and Testing Hypotheses with the Data

This page authored by Alfred P. Montero, Carleton College.
Author Profile
This material was developed as part of the Carleton Teaching Activity Collection and is replicated on a number of sites as part of the SERC Pedagogic Service Project


This project is composed of two sequential parts: a six-week process of building two datasets (a cross-sectional and a time-series, cross-sectional dataset) containing electoral, party system, and voting behavior data for a set of countries and a three-week project employing non-parametric and regression techniques to test established hypotheses in the democratic institutional literature as well as hypotheses of the students' own design.

Learning Goals

The pedagogical goals of this term-long project include advancing skills at the middle-division level in data-gathering, statistical analysis, hypothesis-testing, and oral as well as written communication of findings. Previous knowledge of statistical techniques is not required as these skills will be delivered in the course as part of the piecemeal process of building datasets and conducting non-parametric and regression analysis of the data. The pedagogy of this course favors hand-on active learning techniques in the classroom, the computer lab, and the library.

Context for Use

This project is for use with undergraduates in middle-division political science courses with a focus on formal democratic institutions. The professor must be able to teach methods in the context of the substantive focus of the course on elections, party systems, and mass behavior. Students must learn applications in Excel and commonly used statistical packages such as Stata and SPSS. Although this project is designed to consume an entire course on a trimester system, it may be used in semester-long courses by providing students more time on the separate project components or integrating other dimensions of formal institutional design into the data gathering and analysis components.

Description and Teaching Materials

The two components of the project are discussed in detail in two separate handouts: the Electoral Datasets Project and the Statistical Analysis Project. Additionally, examples of the codebook and the time-series cross-sectional dataset are available. These are the products of a smaller experiment with a similar project in an advanced seminar on time-series cross-sectional analysis of formal democratic institutions.
Assignment handout #1: The Electoral Datasets Project (Acrobat (PDF) 30kB Oct28 08)
Assignment handout #2: The Statistical Analysis Project (Acrobat (PDF) 7kB Oct28 08)
Codebook for TSCS Dataset Composed by Students, Fall 2006 (Microsoft Word 35kB Oct28 08)
TSCS Dataset Composed by Students, Fall 2006 (Excel 444kB Oct28 08)

Teaching Notes and Tips

Depending on the desired scope of the statistical analysis, the professor may opt to add socio-economic, demographic, and other variables to the dataset project to test a wider range of hypotheses. The TSCS dataset example from fall 2006 included a range of dummy variables as well as socio-economic variables for testing.


The assessment tools for this project are eclectic. Constant communication among the student research teams and between the teams/individual students and the professor is essential. On-line discussion or chat forums, available in course management systems such as Moodle are effective in this regard. Team meetings during professor office hours are also useful. Meetings in the library during the data gathering segments and meetings in the computer lab during dataset composing and analysis segments are also extremely useful. All of these techniques enhance the ability of the professor to oversee what students are doing and whether they are learning the key lessons from each task.

References and Resources

Full resources for this project are discussed in the first handout on the project.