Spring 2000


Professor Robert B. Packer
Office: Dept. of Political Science, 413 Willis
Phone: 646-4119
e-mail: rpacker@carleton.edu

As the title implies, Political Science 190 introduces students to scientific inquiry in political science research. Specifically, we survey quantitative data collection and basic methods of statistical analysis. Emphasis is placed on the analysis of data and the substantive interpretation of results. Of necessity, students will be exposed to a variety of algebraic formulae by way of illustrating some of the more commonly used statistical procedures, but the math will be kept to a minimum. Our focus will be an intuitive understanding of quantitative research design and application. Those desiring more rigorous mathematical treatments are encouraged to take follow-up courses in statistical theory and econometrics.

This course is designed to evaluate three general approaches to research, case studies, correlational analysis, and formal theory, to sharpen our understanding of the strengths and weaknesses of each so we can make more informed choices about using them in our own research and evaluating their use by others. Each major approach will then be analyzed separately, looking at the arguments about its strengths and weaknesses, some examples of its use taken from the current literature. We will conclude with a brief discussion of phiolosophy of science. Assignments include two abstracts of data-based articles - each counting for 10% of grade. There is an assignment on evaluating a model - a thorough critique answering the twelve questions posed by Johnson & Joslyn on pp. 403-404 - as well of class exercises.


Abstracts (10% each)


Lave & March exercises


Scientific method exercises


Paper on modeling


Summary paper


Class participation


Course Outline

I. Introduction to Modeling

March 29-31
Singleton, Straits, & Straits, Approaches to Social Research, 2nd ed. (1993): 40-63.

Lave & March, An Introduction to Models in the Social Sciences (1975): 10-84.

Assignment: write responses to problems 2, 3, and 12 in chapter 2 and problems 3 and 4 in chapter 3

II. The Scientific Method

April 3-5
Bremer, Cannizzo, Kegley, & Ray, "The Scientific Study of War: A Learning Package," in Vasquez, The Scientific Study of Peace and War (1992): 373-437.

Assignment: write responses to exercises in the learning package

III. Modeling

April 7
Johnson & Joslyn, Political Science Research Methods, 3rd ed. (1995): 1-18, 41-71, 153-169.

April 10: typology

Krasner, "State Power and the Structure of International Trade," in Frieden & Lake, International Political Economy: Perspectives on Global Power and Wealth, 3rd ed. (1995): 19-36.

April 12: two-by-two contingency tables

Wallace, "Armaments and Escalation: Two Competing Hypotheses" in Vasquez, The Scientific Study of Peace and War (1992): 75-92.

Diehl, "Arms Races and Escalation: A Closer Look" in Vasquez, The Scientific Study of Peace and War (1992): 93-108.

April 14: correlational analysis

Levy, "Alliance Formation and War Behavior: An Analysis of Great Powers, 1495-1975," in Vasquez, The Scientific Study of Peace and War (1992): 3-36.

Wayman, "Bipolarity and War: The Role of Capability Concentration and Alliance Patterns among Major Powers, 1816-1965," in Vasquez, The Scientific Study of Peace and War (1992): 177-203.

April 17: case study

Lieberman, "Deterrence Theory: Success or Failure in Arab-Israeli Wars?" [Part 1: Designing Around: The "War of Attrition," Success or Failure?]

April 19: surveys

Johnson & Joslyn, Political Science Research Methods, 3rd ed. (1995): 375-401.

Schlozman, Bums, & Verba, "Gender and Pathways to Participation: The Role of Resources"

Wilcox, "Race Differences in Abortion Attitudes"

April 21: experiments

Johnson & Joslyn, Political Science Research Methods, 3rd ed. (1995): 111-152, 197-259.

April 24: expected utility and game theory

Bueno de Mesquita, Principles of International Politics (2000): 22-55.

IV. Philosophy of Science

April 26: behaviorial revolution
Johnson & Joslyn, Political Science Research Methods, 3rd ed. (1995): 19-39.

Easton, "The Current Meaning of Behaviorialism in Political Science"

Sibley, "The Limitations of Behaviorialism"

April 28: rival methods

Miller, "Postivism, Historicism, and Political Inquiry"

R. B. Packer
POSC 320
Abstract 1

Ravenhill, John. "Comparing Regime Performance in Africa:

the limitations of Cross-National Aggregate Analysis," The Journal of Modern African Studies, 18, 1 (1980): 99-126.

Query: Are military regimes more effective in "modernizing" and implementing "national development" strategies than civilian regimes in Africa?

Spatial temporal domain: 50 regimes from 33 countries, plus 14 "mixed regimes" in sub-Saharan Africa for years 1960-73.


Outcome: economic performance, socioeconomic performance, military spending.

Predictor variable: regime type.

Data Operations: Economic performance is measured in terms of growth rates in constant gross national product, constant gross domestic investment, constant exports, and international reserves. Socioeconomic performance is assessed by comparing rates of growth of primary-school enrollment and of the food-price index. military spending is measured by constant military expenditure and military expenditure as a percentage of GNP. Regime type is coded by authors as "civilian," "military," and "mixed."

Data Sources: Data for GNP, GDI, and exports were taken from I.B.R.D. World Tables, 1976, international reserves were derived from U.N. Statistical Yearbook. Socioeconomic data were taken from the UN Economic Commission for Africa, African Statistical Yearbook, UNESCO Statistical Yearbook, and UN Statistical Yearbook. Military expenditure data taken from World military Expenditures and Arms Trade. 1963-73 and 1965-75.

Data Manipulation: Data are assembled to enable a cross-sectional comparison of 19 civilian regimes (uninterrupted for 13 year period) and 17 civilian regimes both prior and/or after a coup d'etat with 14 military regimes (for any time during 13 year period); there is a "control group" (i.e., "mixed regimes") for all regimes in countries subject to military rule

Data Analysis: The principal statistical technique utilized for all sets of data is analysis of variance: outcome variable by regime type. Analysis of covariance, controlling for GNP, was also employed in order to examine the relationship of initial levels of development to regime performance and military expenditure. As a means of checking the accuracy of results, multiple regression with a dummy variable for regime type was replicated with the data.

Findings: There are significant correlations between the four economic performance variables, although the growth in international reserves is related significantly only to the growth of exports. Low or negative correlations are reported between the rates of growth of GNP and GDI, and of primary-school enrollment and the food-price index. But significant correlations occur between the rate of growth of the food-price index and those of international reserves, military expenditures, and military expenditure as a percentage of GNP.

One-way analysis of variance on the performance variables confirms the expectation that there are no significant differences between the aggregate performances of the four regime types on any of the outcome variables. While there are substantial differences between the mean scores for the four groups on all variables except the rate of growth of exports, the individual scores are so widely dispersed that the 95 percent confidence interval around each of the mean scores includes those of all other groups. No evidence whatsoever is found to support the conclusion of other scholars that military rule in Africa has a positive effect on the rate of growth of GNP. In fact, the mean score for the military group on this variable was the lowest of the four types of regimes.

Further analysis by way of T-tests of the differences between group means shows significant contrasts on only two of the variables. Countries with uninterrupted civilian rule have significantly lower (at the 0.05 level) rates of growth in primary-school enrollment than civilian regimes in countries where coups have occurred. This may be due to the f act that most such regimes ruled immediately af ter independence when rapid expansion of the primary-school enrollment occurred on a numerically small base. Civilian regimes record a significantly lower rate of growth (at the 0.05 level) that military regimes on the food-price index variable. Again, the temporal dimension may be of importance. The mean growth in the cost-of-living for civilian regimes is substantially below that for the control "mixed" regimes.

Mean rates of growth for four states lacking armed forces (i.e. Botswana, The Gambia, Lesotho, and Swaziland) were significantly higher than those for civilian regimes prior to/after military rule, military regimes, and the control "mixed" regimes categories on the GNP variable, and significantly higher on GDI than other civilian regimes and civilian regimes prior to/after military rule (at the 0.05 level) and military regimes and the control "mixed" regimes (at the 0.01 level). However, the rate of growth of primary enrollment was significantly lower (at the 0.05 level) than that of the civilian regimes prior to/after military rule -- again probably a function of the later period under consideration.