Seminar in American Government:
Interest Representation in American Politics
E. Uslaner Fall 2001
Tydings 2126E x54151
This is the "core" seminar in American politics, designed to provide a survey of the most important literature in the field. Nevertheless, the course does have an overall theme: interest representation in the American political system. Who benefits and who loses from decisions made in American politics, especially at the national level? How democratic can the country be? How democratic should it be? What is the role of various institutions (the legislature, executive branch, the courts, political parties, the ballot box) in responding to diverse interests? Can the government resolve complex policy issues while also being responsive to these diverse interest? Did the founders of the country intend the government to do so?
As we examine these questions, we shall first concentrate on some theoretical issues and then move on to a more detailed analysis of specific institutions. At all times, we shall try to strike a balance between maintaining a consistent theme throughout the seminar and the necessity to cover the most important literature on a wide variety of subjects within the scope of American politics.
Course grades will be based upon: (1) class participation (20 percent); (2) the short paper analyzing the hypotheses in the readings (20 percent); (3) the initial 3-page research design due in Week 4 (10 percent); (3) a research design or completed paper on some topic in American politics of not less than 25 pages (40 percent); and (4) the oral presentation of the research design (10 percent).
Each week a seminar members will lead the discussion by examining a major hypothesis in the week's readings. Do the readings agree on the importance of the hypothesis? Do they agree on the hypothesis itself? What is the justification for the hypothesis? How strong is that justification? Could you offer an alternative thesis that is equally plausible? Does another reading of that week offer a contrary hypothesis? Which seems more likely to be confirmed? Why? Suggest a research design that would allow us to test this hypothesis against a reasonable (better yet, a prominent) alternative.
Students will present arguments to the class and also summarize them in a short paper (approximately 5 pages).
The research design asks you to state your own hypothesis, ground it theoretically in a literature, and present a way of testing your hypothesis. The hypothesis does NOT have to be based upon quantitative data. Case studies are quite appropriate. The essential point is that you derive your hypothesis from an important question and that you justify it theoretically. Whether your data are qualitative or quantitative, your method of testing your hypothesis should be appropriate to your question. Pick a question that interests you. This should make it easier for you to interest others (including me). You will have to convince others--and soon. You must have a preliminary hypothesis by Week 3. Beginning in Week 4, three seminar members per class will present their hypotheses to the class for critical examination. AT WEEK 4, YOU MUST HAND IN A VERY SHORT (ABOUT THREE PAGES) PAPER STATING YOUR HYPOTHESIS AND JUSTIFYING IT. On December 4 and 11, students will present their papers to the seminar.
All reading below is required; it should all be on reserve at McKeldin. Those books marked by an asterisk should be purchased. They are available at the various area bookstores (University Book Center, Maryland Book Exchange, etc.). For those who may want to economize somewhat, you might consider the library as an alternative reading source. Serious candidates in the Ph.D. program would probably want to have all of the required books in their libraries. There is also a packet of xeroxed materials available at Bel-Jean's.
1. (9/11) The Roots of Politics
*Schattschneider, Semisovereign People, entire.
*Mansbridge, How We Lost the ERA,entire
2. American Political Culture and Participation
*Tocqueville, Democracy in America, v. 2, entire
Sven Steinmo, "American Exceptionalism Reconsidered: Culture or Institutions?" in Lawrence C. Dodd and Calvin Jillson, eds., The Dynamics of American Politics (Boulder, CO: Westview, 1994), pp. 106-131.
Rogers Smith, "Beyond Tocqueville, Myrdal, and Hartz: The Multiple Traditions in America," American Political Science Review, v. 87 (September 1993), pp. 549-566.
This class needs to be rescheduled for Friday 9/21; no class 9/25
3. (10/2) Civic Participation in America
*Verba, Schlozman, and Brady, Voice and Equality
Robert D. Putnam, "Bowling Alone," Journal of Democracy, v. 6 (January, 1995), pp. 65-78.
4. (10/9) American Values in Theory and Practice
*Hochschild, Facing Up to the American Dream, chs. 1-7, 12-14
*Dawson, Behind the Mule, entire
5. (10/16) How Interests Get Represented in the United States
*Walker, Mobilizing Interest Groups in America, entire.
*Dahl, Who Governs?, chs. 1-12, 16-28.
6. (10/23) Parties and Interest Representation: Tyrannizing Whom?
*Hofstadter, The Idea of a Party System, entire
*Aldrich, Why Parties?, entire
7. (10/30) The Presidency: Too Much or Too Little Power?
*Lowi, The Personal President, entire.
Skowronek, "Notes on the Presidency in the Political Order," pp. 286-302 in Orren and Skowronek, Studies in American Political Development, v. 1.
8. (11/6) The First Branch of Government: Responsive, Bifurcated, or Both?
*Wilson, Congressional Government, entire
*Mayhew, Divided We Govern, entire.
9. (11/13) The Courts: Savior, Intruder, or Neither?
*Rosenberg, The Hollow Hope, entire
Graber, "The Nonmajoritarian Difficulty: Legislative Deference to the Judiciary," in Orren and Skowronek, Studies in American Political Development, v. 7, 35-73.
10. (11/20) The Accountable Electorate: Is Representation a Myth?
*Campbell et al., American Voter, chs. 1-7, 9, 19-20.
*Downs, Economic Theory of Democracy, chs. 1-8, 11-14.
11. (11/27) Linkage Models of Public Opinion
*Page and Shapiro, The Rational Public, entire.
*Ginsberg, The Captive Public, chs. 1 and 3-4.
12/4, 12/11 Student Paper Presentations
WRITING A RESEARCH DESIGN
First, THINK SMALL. Think of a research question you would like to answer. Or think of an explanation that you find unsatisfactory. Thinking small means you should be specific in posing a hypothesis.
A hypothesis is a statement of the form:
The more X, the more Y.
The more X, the less Y.
Y is your dependent variable, what you want to explain. X is your independent variable--the factor that "explains" your dependent variable. Don't take this too far, but you might think about them in terms of cause (independent variable) and effect (dependent variable). Once you have taken GVPT 622 and especially GVPT 700, you will understand why cause-and-effect statements are not appropriate for the data we have, but for now the analogy will do.
Thinking small means you must be specific. A hypothesis is NOT an explanation of how the world works. Don't start off by writing that you want to understand the bases of Presidential power or why Congress is corrupt.
Instead, consider the following narrower hypotheses that lend themselves to testing:
1. The greater the rate of inflation, the more seats the President's party will lose in midterm Congressional elections.
2. The higher a President's popularity, the more likely the President will be to enact his/her program in Congress.
3. States with individualistic political culture will be more likely to have high levels of welfare spending than states with moralistic or traditional cultures.
4. Voters with the highest level of political efficacy are most likely to vote for incumbent members of Congress.
These four hypotheses may be true or false. That doesn't matter. What is key is that you can formulate your hypothesis clearly and justify it. Justification is key. Are all four hypotheses equally plausible?
To get an idea of plausibility, you need to consider two fundamental issues. First, what do the components of the hypothesis mean? Second, what is the logic behind your hypothesis?
Consider hypothesis 1. It is the most straightforward since the dependent and independent variables are clear. Your Y, the dependent variable, is how many seats a President's party will lose in midterm Congressional elections. It is straightforward to measure this. Your X, the independent variable, is the rate of inflation, which has a standard measure available for a variety of years from The Economic Report of the President.
Hypotheses 2, 3, and 4 are more complex. How do you measure (hypothesis 2) the level of Presidential support in Congress? A variety of measures are available and you will have to choose among them. What are individualistic, traditionalistic, and moralistic political cultures and why should they affect policy-making in the American states (hypothesis 3)? How might we measure "political culture" anyway? Is culture important or is it a surrogate for something else? What is political efficacy and how might we measure it (hypothesis 4)?
Figuring out how to measure your variables is essential, but it is secondary to justifying your hypothesis. You MUST have a hypothesis that makes sense. Your hypothesis should first and foremost be interesting to people beyond yourself. Pick a topic that has the potential to grab others' attention. If the topic is exciting to begin with, so much the better. If it is not, it is your task to convince others why your topic matters. Some topics are simply too boring to test: Liberals prefer more social welfare spending than conservatives. Others are silly: Tall people are more likely to be highly educated than short people. Pick a topic that is interesting because testing it would resolve an intellectual puzzle.
Once you make the case that your topic is worthwhile, you need to justify it. Consider the following fascinating hypothesis (which is in fact true):
The larger the number of storks in counties in Finland, the higher the county's birthrate.
It would be remarkable if storks did bring babies, but do you believe it?
Where do you get your hypothesis? Think of topics that interest you. If you're not fascinated by your topic, nobody else will be. Think of a question you want to answer--a puzzle that grabs YOU. Where do you get ideas? From your own thinking, news stories that fascinate you, or from your own reading. Is there an intellectual puzzle that you find exciting or a result that you have read that you find unsatisfactory? If the latter, why are you not convinced by someone else's argument? Is there a better way to think about the problem?
After you have come up with your hypothesis, you need to shift gears. Now, you must THINK BIG. You must link your hypothesis with a larger body of literature. Many articles and books call this a "literature review." I don't like literature reviews. They tend to be too discursive, covering every piece of research that touches on your hypothesis, whether directly relevant or not. Instead of a literature review, you should look for literature that bears directly on your hypothesis. Instead of a "literature review," you should write a "theoretical perspectives" section in which you examine literature specifically relating to your hypothesis. To get there, you must begin as if you were doing a literature review.
If you want to examine the impact of inflation on midterm voting on Congressional elections, you should consult the literatures on voting behavior more generally, on economics and politics, and on Congressional elections. These are vast literatures and will tell you much more than you need to know to test your specific hypothesis. Cull through these literatures to determine which studies specifically deal with the impact of economic factors on Congressional elections. What factors lead people to vote the way they do in Congressional elections? How are midterms different from Presidential year contests? Do people hold the President responsible for economic outcomes? (The answer to this may be less straightforward than you think.) Is inflation the economic factor most likely to have an impact on elections (what other factors might be relevant)? Does the President's party generally lose seats in midterm contests? Might factors other than the economy lead to seat losses when they occur?
At the end of your theoretical section, you should come up with either: (1) a general framework justifying your hypothesis; or (2) two (no more than two) alternative frameworks, one that supports your thesis and another that denies it. Under either scenario, you should meld the diverse literatures into a coherent argument supporting your thesis. You might also consider an alternative argument rejecting your hypothesis, if there are clear lines of division in the literature you are reviewing. WHAT YOU SHOULD AVOID IS CITING A LOT OF LITERATURE JUST TO BE INCLUSIVE. CITE WHAT YOU NEED FOR YOUR THESIS; DON'T ADD A LOT OF FLUFF SO THAT YOU CAN HAVE A LONG CITATION LIST.
To this point, you have selected a topic, indicated why it is (or should be) intereresting to others), and justified it by drawing on existing literature in one or more fields. The next step is to figure out your methodology.
Methodology includes determining your data base and the appropriate technique for analyzing your data. You need not feel restricted to quantitative techniques. If you employ case studies you must show why the cases you select are appropriate and how broad your conclusions can be. Dahl's Who Governs? is designed to show that pluralist politics reigns in American cities. Is New Haven typical? Might other cities--or a different method--show that a small "power elite" actually controls decision making?
Let's go back to the inflation-Congressional elections example. You decide to collect data on seat losses for the President's party and on the inflation rate. So far, so good. But for how many years? If you only use 10 years, your statistical results will be rather weak (10 is a rather small number of cases, particularly if you can gather data for a larger time period). So suppose you decide to collect data for the entire 19th and 20th centuries. You will find that economic data for much of the 19th century are either unavailable or unreliable. So you decide to limit your analysis to the 20th century. You still must ask yourself: Did the basis of voting behavior change at all during this time? The 1930s witnessed a massive political realignment. The 1970s began a period of dealignment in which voters became less attached to political parties and more likely to support incumbents. As a result, they were less likely to punish a President's party for the economic results that a President "produced." You need to be aware of these changes and to incorporate them into your analysis.
Once you have determined your data base, you need to select your method of analysis. For hypothesis 1, you select regression analysis (which you will learn about in GVPT 622). Regression is a statistical technique allowing you to test the hypothesis: The more X, the more Y. Regression analysis takes a form that should be familiar to you from high school algebra (assuming your memory is good):
Y = a + bX,
which is the formula for a straight line. Y is your dependent variable, X your independent variable. a is the intercept on the Y axis that indicates the value of Y when X = 0. Here, it would tell you how many seats the President's party would lose if the inflation rate were zero. b is the "slope," telling you the impact of X on Y. If b=1 and you measure inflation and seat losses for the President as percentages, you would conclude that:
for each one percent rise in the inflation rate, the President's party will lose one percent of its seats in the next midterm election.
If the regression equation were:
Y = .2 + 1.2X,
you would conclude that for each 1% change in the inflation rate, a President's party would lose 1.2% of its seats in the next election plus .2% of its seats that it would lose because of other factors (the value of the intercept equals the seat loss when the inflation level is zero, so you must attribute the additional seat loss to other factors).
Again, so far, so good. But you still need to go further. Is the inflation rate the only factor that affects seat loss? Might the result be spurious? What is a spurious result? Think of the example of the birth rate in Finnish counties. It is well supported statistically, but it makes no sense. To test for a spurious relationship, we need to employ control variables. A reasonable control variable for our Finnish example is how rural each county is. So we wonder whether rural counties have both large stork populations (you would be no more likely to see a stork on the streets of Helsinki than you would be to see one in the District) and higher birth rates. So you reformuate your regression equation:
Birth rate= a + b1*(Stork Population) + b2*(Rural Percent)
When you test this hypothesis, you find that the impact of the stork population vanishes once you control for how rural a county is.
For hypothesis 1, you might control for the following factors: (1) Presidential popularity (when data are available); (2) how many seats the President's party controls in Congress; and (3) the party of the President. Why would you want to control for these variables? Are there other variables you might suggest?
You now have specified your data base--and where you can get your data (published sources, your own survey, your investigation of specific cases)--and your methodology. Now you should tell us what you would conclude if your hypothesis were true. Are there implications for the theoretical perspectives you considered? Are there any public policy implications? If you found support for hypothesis 1, does this imply that politics ultimately depends upon economics? If so, what does this tell us about theories of voting behavior? Is this good or bad? Should we all apply to the economics department and give up political science--or should economists all join us? What are the limitations of your analysis? The data you have gathered are all "aggregate" data (the total seat loss for each year). They tell us nothing about how individual voters behave. How might we go beyond your study to examine voters' reactions to inflation and other factors?
If your hypothesis does not work out, why do you think it did not receive support? Might you suggest some alternative ideas as to what "causes" your dependent variable? If your hypothesis didn't work out, what would the implications be for the theoretical perspectives you considered when you justified your hypothesis?
In your research design, you should consider the implications for your theoretical perspective(s) under both scenarios: Your hypothesis works out or it doesn't.
PREPARING A RESEARCH DESIGN FROM SOMEONE ELSE'S WORK
Follow the overall logic in the above section, "Writing a Research Design." You should read the book/article trying to reconstruct the author's research design.
Your logic should be:
1. What is the problem being investigated? Is it important or interesting?
2. What is the author's central hypothesis? (Many articles and every book will have many hypotheses. Select ONE for your paper.)
3. Justify the hypothesis briefly (2-3 pages) as the author has done.
4. Describe the data employed and the methodology employed to test the central hypothesis.
5. What are the implications of the test of the central hypothesis for the author's theoretical perspective.
6. Are you convinced that the author's hypothesis is reasonable, that the justification makes sense, that the data and methodology are appropriate, and that the author makes fitting inferences about the conclusions? If so, state why. If not, why not?