No CrossRef data available.
Article contents
Interactive Data Analysis of the 1980 Election
Published online by Cambridge University Press: 12 May 2020
Extract
By introducing data analysis techniques, analyzing current research in voting behavior, and using the computer for social science research, introductory American Government courses on the undergraduate level are undertaking ambitious goals. Using the SETUPS units for each presidential election since 1972, cadets at the Coast Guard Academy have engaged in a comprehensive approach to the study of American electoral behavior in an interactive data analysis mode. Although SETUPS modules are run at most colleges and universities in batch mode, there are imitations and drawbacks to batch mode. The Coast Guard Academy is able to utilize interactive data analysis as a result of its time-sharing arrangement with the Dartmouth College Time Sharing System (DCTS). In this article we will discuss the use of SETUPS in an interactive mode using the CIDAR interactive data analysis system.
- Type
- Research Article
- Information
- Copyright
- Copyright © American Political Science Association 1983
References
Notes
1. SETUPS was initially taught using the IMPRESS system, but it became obvious, with time, that this system had limitations that interfered with the way the staff wanted to teach SETUPS. Given the large number of cases in the election study SETUPS, IMPRESS would allow the analysis of only five variables at a time. Before beginning analysis, a teacher or student had to select which five variables to analyze. To change one of the variables or to add a new one, the IMPRESS run had to be stopped and begun again. The recoding or mapping of variables had to preceed the analysis, which made the system awkward. Although the ability to carry out some simple data transformations is available, IMPRESS does not permit index construction or other mathematical operations in the middle of a run. Nor can one select specific cases for processing after the run begins. IMPRESS also tends to refuse to print the statistics for contingency tables if the cell values are less than five. While it is statistically correct to make corrections in chi square for cell sizes less than five, it is difficult, from a teachers point of view, not to be able to examine the tables. To correct the problem, the teacher has to stop the system, restart it, and group the data before coming to the table.
2. The computer project assignment sheet is as follows.
Your assignment, if you wish to accept it (and you will because it comprises 15 percent of the term grade) is to do two cross-tabulations. The first cross-tabulation will use any two variables from the codebook from the SETUPS Voting Behavior: The 1980 Election. The selection cross-tabulation will use the original two variables controlling for a third. The requirements are a 3 to 5 page write-up in the following format:
I. Introduction
What study did you use? What substantive question did you investigate? State this in terms of concepts (e.g. educational background) and not in terms of variables (e.g. EDUCAT or V24).
Why did you select the variables? Which is the independent variable? Which is the dependent variable? Why?
II. Findings
What important association did you find? Be certain to explain clearly what your tables show.
What variables had no association? Explain the values for chi square and Yule's Q or Gamma. Be sure to include your printout with the paper.
III. Conclusions
What is the meaning of your data? Be sure to include a copy of the tables you are analyzing within the paper.