Sign the following pledge and staple this sheet to the front of your paper.
I have neither given nor received any unauthorized aid in writing this research report. I understand that violating this pledge means failure in the course and possible expulsion from the university.
Signature: ______________________________
Instructions:
The purpose of this research paper is to bring together all of the material (or nearly all of it) you have learned during the semester in a practical and meaningful way. What you have learned should enable you to perform a variety of studies that are scientifically valid and reliable after you leave your formal education. In other words, you have a practical skill you can sell. You should get a sense of "WOW! This all seems to fit together now!" At least I hope so.
We began class the first day of the semester talking about the steps of scientific research and then spent the rest of the semester tearing down each step into practical exercises. Your write-up should follow the same steps. Let me review these steps in outline form and list the questions you should answer in each step. YOU WOULD BE VERY VERY WISE TO USE THESE STEPS AS HEADINGS IN YOUR PAPER--AN OUTLINE AROUND WHICH YOUR PAPER IS BUILT. That will make it easy for me to read and easy for you to see if you have covered all the things you need to cover. If you don't answer the questions I pose to you under each of these steps, you will lose points. Need I say more on this subject?
The paper should be typed (though you can do the tables by hand or cut and tape them into the text), double spaced, 12 point font. I want the tables/graphs in the text where you talk about them—DO NOT PUT THEM AT THE END!
Here is an outline of what you should have in your paper. Hopefully most of it is already there from what we have been doing in class.
I. Problem Statement -- Here you should describe
the problem or question area you chose to study. Focus is usually primarily on the dependent
variable. Why is it important? What
general public policy implications does it have? Or why is it important
to political science or whatever? Most of you have already written a good
draft on this, so start with that and improve it a little if you can.
The independent variable can be important if understanding
why and how it affects the dependent variable has important policy or political
implications.
II. Theory and Concepts -- it's probably easiest to combine these two steps, or you can do them as two separate steps.
You can talk about either one first, though theory usually comes first. The names of labels for the concepts will have to be stated in the theory.
Theory is where you really need some references in the literature (again, standard texts are your best bet for starting out -- then you look up articles that they reference). What have previous researchers found.? Do they suggest relationships that you can test? Perhaps you will expect new relationships than in the past because of the nature of the times we live in or because of the current generation of students. Remember that any survey is a snapshot in time!
Then you move to describe and define the concepts involved in your problem statement (like education, civil liberties, political efficacy, political trust, partisanship, media consumption, ideology, issue positions, generational groups, regional id, and so on). You should be able to do that in several short sentences. Where possible, use a text book definition (which is more likely to be "appropriate") and reference the text from which you got it, or show that you are using a definition that is the same as in some other scholarly work. Dictionaries of political analysis can be helpful here. American government texts and public opinion texts can be quite useful here as well. Remember that some concepts that may seem simple like age are not really the concept at all. Age may simply be a measuer of world experience or maturity, the real concepts! The same can be said of year in school, which is just a measure of the amount to college experience.
Remember that general theories are better than specific ones. For example, if you are talking about someone adopting an opinion because it should be in the self-interest of some group in which that person is a member (like a religious group), then your best theory is based on the psychology of opinion holding. See the references on p. 52 of the 6th edition of Kent Tedin and Robert Erikson, American Public Opinion (New York: Longman). That page also has references to works that explain why people do not always have opinions that are in their self-interest as well. Texts are a great starting place!!
WARNING! You must clearly show that your sources support your ideas. Do NOT just give me some general quotes from sources that have something to do with the variables in your hypothesis and expect me to figure out what those quotations mean. I expect you to do some critical evaluation of the sources to figure out what they suggest about your problem and theory.
III. Hypothesis -- formally state your initial
bivariate hypothesis. State which are the independent and dependent variables
and draw an arrow diagram (or path diagram) to depict the relationships
you are going to test. Then draw a tentative arrow diagram with the confounding
(spurious) and/or conditioning and/or intervening variables included.
A. Bivariate hypothesis: (Example) Among Aiken County residents, those who identify with the Republican Party are more likely to give Bush high job approval ratings and those who identify with the Democratic Party are more likely to give Bush lower ratings.
B. Controls: Next talk about control variables that you think may play a confounding or a conditioning or an intervening role. You should have at least one, and possibly several, depending on what is available. Why are you going to add a control variable and why did you choose the particular control variable you did? That is, why should it be related to the other two concepts you have already chosen for the independent and dependent variables? Why and how might it change the bivariate relationship? This is again theoretical thinking. It could be put in the theory section just as easily.
Once you have talked about variables that will help you establish support for some kind of causal relationship, discuss variables that may act as intervening variables. You don’t have to have this unless there is an obvious one that we measured in the survey or if you had no possible confounding or conditioning variables you could use. Confounding variables are certainly the most important ones to consider since they can cause the bivariate hypothesis to become spurious!
Example of a possible conditoning variable for a general population survey: Being troubled by the costs of the war in Iraq could affect the relationship between partisanship and Bush's job approval. For those who feel that the war is not worth it, one might expect that the relationship will disappear, while those who either approve or have mixed feelings should have the expected relationship between partisan identification and Bush job approval.
Party ID à Pres Job Approval Rating
/\
|
War in Iraq position
IV. Operationalization – Here you explain how each of the variables you will be using was measured. You must give the exact wording in quotation marks of the questions used and the answers that were possible.
You must also explain how you collapsed variables into groups or combined variables to create new compound variables.
Make sure you discuss the validity and reliability of each question used. (Pretesting with focus groups and using standard questions--as we did with the class--add to both reliability and validity. Alternatively, using well used standard questions allows you to assume that the questions are both reliable and valid. Using questions from national surveys also helps here. But as you saw with some of the national questions, some of them may not be very reliable because they are often too long!) In this discussion, you might be wise to state what validity and reliability mean.
V. Data Gathering (Methodology)—hopefully you already have this written. It should include: population, type of survey, how the sample was chosen (I want some detail here--the things we did to minimize sample bias), how interviewers were trained (if that is relevant--it is not with self-administered online surveys), why we trained them--to reduce interviewer bias), when/where interviews were performed, how supervised, response rate and how we tried to maximize the response rate, sample size, and sampling error (calculate this for questions that involve all the interviews performed).
VI. Analysis -- Present the bivariate table/graph with all variables and codes clearly labeled, discuss the table (right below it, where you say things like "As we move from..."), present the chi square (if appropriate) and significance level of that chi square (if you have any trend or shift), and discuss what the significance (p=???) tells the reader. ("There is a ?% chance that this relationship could exist in the sample if no relationship existed in the population.") If you have two ratio level variables, you will have a regression with a scatterplot here instead of a table. If you have one interval and one nominal variable, you will have an analysis of variance (ANOVA) here. Discuss statistical significance for any test you do. Then you should discuss the strength of the relationship using the appropriate statistical measure.
Present the control tables (or regressions or analyses of variance) for each possible value of the control variable. Use with the same discussion after each. Interpret each table in relation to the bivariate relationship--what effect did it have, if any, on the original bivariate relationship? Account for significance and strength. After you do this for all values of the control variable, you should be able to draw some conclusion about that the actual relationship is among all three of your variables. You may have to redraw the path diagram.
You
can copy and paste charts, tables, graphs them from MicroCase. Please eliminate the Missing data row
and column as those numbers only confuse the reader. Only include the statistics
that are appropriate. Put the tables in the proper place--I don't want to have to flip to the end
of the paper to look at them!
Make sure everything in a table lines up as it is supposed
to--pasting from MicroCase to Word does not always do that.
VII. Theory Reformulation—Restate your
original problem and what you thought theory would tell us about the problem.
review your findings from the bivariate relationship and how the control
affected the relationship. Then tell me what this tells us about the original theory. Do the results support
the theory? Do they fail to support it? Do they support it only under certain
conditions? How does the independent variable actually affect the dependent
variable? Here is where a possible intervening variable adds to your explanation.
Do your findings suggest a new line of additional research? If so, what would
you look at next if you wanted to continue this research? Have conditions
changed so that one can expect new relationships in a new political environment.
Be thoughtful and creative here.
-- "Works Cited," or References. You should have a minimum of 5 separate scholarly references, listed at the end as "Works Cited." Some of the references may be here in the problem statement section and some may be in the next section on theory—you had better have at least one for theory!! Remember that a minimum is just that -- a minimum. More helps earn you a better than minimum grade! They should all be in APA format!
That's it!!! You will no doubt need some help. Get it FROM ME AND NO ONE ELSE, although I will authorize you to ask other students routine questions about MicroCase. I'd be happy to comment on your ideas and graphs or charts.
It must be typed in 12 point font with one inch margins and double spaced. Grammar and spelling and punctuation will count!