I. Introduction

II. Intro to preparing to write your papers. Questions you should be able to answer, or at least thinking about by now

A. What is your dependent variable (DV), i.e., what are you trying to explain? For example, are you trying to explain changes in harvest of fish, emissions of some pollutant, amount of lumber exported?

1. Make sure you have graphed it.

2. Don't have excessively complicated graphs.

3. Have several comparisons on graphs.

B. Have you developed a list of factors (independent variables or IVs) that might explain change in your DV? This list should consist of two types of variables:

1. NON-treaty variables -- imagine there hadn't been a treaty at all -- what would you use to explain the variation in the DV that you observe? These are your "alternative explanations" or "rival hypotheses."

2. General treaty variables -- membership, area regulated, species/pollutant regulated, etc.

3. What features of the convention are causing it to have an effect or not to have an effect. That is, focus on explaining WHY this convention did, or did not, have the desired effect.

4. For example, sanctions, monitoring, financial provisions, etc. Make sure you can clarify your treaty IV and its expected impact – read the treaty, protocol or convention and find out what it required and when it was likely to have an influence. E.g., LRTAP Protocols

C. What are the bases from which you will develop counterfactuals? Remember, the goal is to have a graph in which you compare the observed regulated behavior of interest (the DV!) to one or more counterfactual lines of what that regulated behavior would have looked like had it not been regulated. Possible bases for generating counterfactuals are:

1. The behavior of MEMBERS BEFORE the treaty. For example, did the United States catch different levels of salmon in the North Atlantic after the North Atlantic Salmon treaty was signed than it did before that treaty was signed?

2. The behavior of NON-MEMBERS AFTER the treaty (and best if the non-members were behaving similarly to the members before the treaty). For example, did Chile (a non-member) catch different levels of salmon in the North Atlantic after the North Atlantic Salmon treaty was signed than did the United States (a member) after that treaty was signed?

3. The same or similar type of behavior by MEMBERS in a location that was not regulated. For example, did the US catch different levels of salmon in the South Atlantic than in the North Atlantic in response to the North Atlantic Salmon treaty?

4. A similar but non-regulated type of behavior by MEMBERS in the area that was regulated. For example, did the US catch different levels of halibut (unregulated) than salmon (regulated) in the North Atlantic in response to the North Atlantic Salmon treaty?

D. Are there possibilities that among member countries, the treaty might have influenced countries differently. For example, were Art. 5 countries influenced differently than non-Art. 5 countries by the Montreal Protocol?

E. Even if the treaty didn't treat countries differently, does it seem like some countries were influenced by the treaty and others weren't. Use Sprinz and Vaahtoranta's article -- perhaps "leader" countries aren't influenced by a treaty (they would have done the right thing anyway) but "laggard" countries were influenced. Or do democratic countries seem to have been more influenced than communist countries?

F. Think about STORY of HOW agreement caused behavior change. OR of HOW other factors caused behavior change.

G. Re-read Mitchell and Bernauer article from reading and follow those steps. – below I treat it as eight steps but both are basically the same.

III. Structure of paper

A. Introduction – written at end, once you know what your argument will be. Make sure you clearly state what you will be arguing.

1. Don’t: "This paper will analyze the Montreal Protocol."

2. Do: "The Montreal Protocol was effective at changing the behavior of developed states but was not effective at changing the behavior of developing states."

B. Definitions and background – keep this as short as possible

C. Theory section

1. What is the general theory you are trying to evaluate? What have authors you have read argued are important features of conventions, treaties or policies that cause them to be effective or not? Are you testing some of these theories?

2. E.g., you might have read that transparency helps improve the effectiveness of a treaty. Write up a section of your paper on why and how transparency helps a treaty be effective. And then, make sure that you discuss how transparent your treaty was (or wasn’t) and how that influenced its effectiveness.

D. Case study

1. Why did you select the countries you did for your study, if you are not looking at the impact on all countries? Why did you compare the two countries you compared? Why did you compare the change over time in the single treaty you studied?

2. Explain exactly what the convention or policy entailed. Explain what impact you would expect it to have, when you would expect it to have that impact, and why you would expect these things. Make sure you are clear on what the expected impact is.

3. Carefully delineate what the independent and dependent variables are, what their values are, and how you think they should vary together. I strongly recommend making a chart like that in the graphing program, with an expected relationship of the IV/DV and the actual relationship you observed.

4. Analyze carefully whether the policy had the intended influence. This should include looking at:

a) Counterfactuals – make sure you make an argument about what would have happened if the convention or policy had not been adopted.

b) Alternative hypotheses – make sure you exclude other explanations of any change in your dependent variable.

E. Generalizing – is what you found likely to be true in other treaties? If so, why? If not, why not – what is unique about this case that made it particularly easy for the treaty to influence behavior.

F. Conclusions –summarize your argument and clarify what the key findings of your paper are.

G. Sources cited – make sure to have a bibliography of the articles and books that you cite in your paper.

1. In doing citations, use the in-text format for citations and the bibliographic style given to you in the Assignment Packet. Don’t waste time trying to create your own or doing footnotes. That will waste your time and not improve your grade.

2. Don’t include articles and books in your bibliography unless you actually cite them. If it wasn’t important enough to cite, than don’t include it.

IV. EIGHT STEPS TO A CONVINCING CAUSAL ARGUMENT

1. Clarify your theoretically-interesting analytic research question.
Begin by clearly stating your analytic question. This should be a question initially motivated by some "puzzle" about why things have turned out the way they have in a particular case. The key is to then move back a step and transform that specific question into a more general, theoretical, and analytic question. Thus, you may be interested in "why did nations only agree to substantial provisions to protect the ozone layer (the Montreal Protocol) in 1987 even though initial scientific links of CFCs to ozone loss were identified in 1974?" This transforms into the more general, theoretical, and analytic question of "Does scientific knowledge have any impact on negotiation of international environmental treaties?"

2. Restate your theoretical question as a simple causal hypothesis.
Simplify your argument into "A causes B" language. Most arguments that you should be trying to make for the purposes of this class can be put in these terms. What is the major independent variable (the cause) and what is the dependent variable (the effect)? In our example, this would be something like "Scientific knowledge of environmental problems has no impact on negotiation of environmental treaties." Remember, of course, that many independent variables may be influencing a specific dependent variable, but that we are attempting to determine which ones really are causes and identify the most important ones.

3. Identify variables, potential values, and possible causal relationships.
Clarify the independent and dependent variables and their potential values. In this case, the independent variable is "Level of scientific knowledge of environmental problem" with potential values of high and low; the dependent variable is "negotiation of international environmental agreement" with potential values of yes and no. Note that you will want to evaluate how variation in the independent variable influences variation in the dependent variable. In this case that means, you want to compare a case in which scientific knowledge of an environmental problem is low with a case in which scientific knowledge is high in order to determine one of four relationships: a) whether an agreement was negotiated in the first case and not in the second case, b) whether an agreement was negotiated in the second case and not in the first case, c) an agreement was negotiated in both cases or d) no agreement was negotiated in either case. In cases c) and d), no causal relationship appears to exist between the independent and dependent variables. In case a), scientific knowledge appears to cause (or at least facilitate) negotiation of an agreement, while in case b), scientific knowledge appears to inhibit signature of an agreement. The goal of your research should be to determine which of these relationships seems to most convincingly reflect what is really happening in the world.
You will also want to pay attention to "control variables," that is, those independent variables that we know (or at least think are likely to) influence the value of the dependent variable but which we want to hold constant so that we can see the effect of the independent variable of interest. In this case, for example, we would want to try to compare two cases in which the US (the hegemonic state) supported negotiation of the agreement, since its failure to support an agreement is likely to strongly influence whether an agreement is signed or not.

4. Look for contrasting cases or generate a counterfactual.
This is similar to the story in the movie "Its a Wonderful Life." To become convinced that A causes B, we must demonstrate that if A were not true, B would also not be true. Here the counterfactual would be that "In the absence of scientific knowledge about the environmental problem, there would NOT have been an international agreement." We can demonstrate this either by finding two cases that are similar in most respects but vary in terms of the level of scientific knowledge and then evaluate whether an agreement was signed or not, or (less convincingly) we can use a "thought-experiment" of saying that "if we imagine a world that was similar to our case in all respects except that scientific knowledge of the problem was low instead of high, there would not have been an agreement."

5. Generate specific predictions and verify argument in absolute terms.
To begin the empirical testing, we need to first generate specific predictions about what the value of our dependent variable should be in the case(s) we are going to study based on the value of our independent variable. Thus, we would want to determine whether the level of scientific knowledge in the ozone case was high or low, and, if it changed over time, when it switched from high to low.
Having predicted the value of the dependent variable, we then can examine the case to determine three things: a) does the independent variable actually have the value we think that it does, "was scientific knowledge (or consensus) really high in the ozone case?", b) what value does the dependent variable actually have, "was an agreement actually signed?", and c) does the counterfactual actually make sense.

6. Develop a convincing causal narrative.
Even if you have shown that the variables correlate, it provides a more convincing and interesting argument if you can tell why and how the independent variable caused change in the dependent variable. For example, you might want to explain that, "as scientists reach a consensus on the cause of an environmental problem and its solution, politicians and negotiators can no longer oppose negotiation of an agreement based on the argument that some scientists do not believe that there is really a problem. Once the majority of scientists agree there is a problem, politicians' interest-based arguments are revealed as interest-based and therefore become less politically powerful." You would then need to provide evidence from the case that this is actually what happened, perhaps showing how the arguments of politicians and negotiators changed as scientists increasingly hound evidence of an environmental problem.

7. Evaluate rival hypotheses that others would argue explain the case better and more simply.
Even if you seem to have made a convincing case for one side or the other of your original thesis, you need to evaluate alternative explanations of the outcomes or effects that we see. For example, others might argue that the reason an agreement was finally reached in 1987 was because Dupont, probably the most powerful CFC producer in the world, finally removed its objections to a treaty for economic reasons. Now, you have to evaluate that argument and show how it is wrong, incomplete or otherwise not adequate to completely destroy your argument that scientific knowledge made a difference. In general, you want to make sure that alternative explanations, also known as "rival hypotheses," do not explain the same outcome better than your theory, that is, do not explain it with fewer and more convincing variables or an otherwise more convincing argument.

8. Evaluate generalizability.
Finally, once you have a convincing argument about the case or cases that you have studied, you want to evaluate whether your argument would also be true about other cases. To what other cases would your argument seem to generalize? Would your argument about scientific knowledge in the ozone case also seem to be true in the cases of whaling, oil pollution, wetlands preservation, acid precipitation? What about cases of arms control, trade negotiations, human rights? Here you will need to go back to your "control variables" and determine to what extent you have limited your generalizability by holding certain variables constant and therefore do not really know how that variation in that control variable actually influences the dependent variable. Essentially here you are trying to determine under what conditions is the theory you have argued for true? Be careful not to claim too much for your theory but, if you think you can make the case, argue why it should be true at least in some other cases.

V. Conclusion