Below are the basic elements a research design should contain. The emphasis given to each element will differ depending on the type of research you are undertaking – e.g., long dissertation versus short journal article – as well as the problem / research question you pose.

For example if what you are primarily doing is updating old research with new data you may spend most of your time on item #4, explaining why your data is far superior to that used in prior studies. If, on the other hand, your argument is that prior research left out key variables you may spend most of your time explaining your model, item #3.

I. BASIC MATERIAL / INTRODUCTION:

This material should appear at the beginning of the research paper. For a book or dissertation, the first 1-2 chapters may be devoted to this material. For a journal article or conference paper it will usually cover the first 1-3 pages.

1. What is your research question?

You begin with a question, not a thesis or purpose statement. (Experienced scholars may begin talking a bout their “research purpose”, but there will be a question embedded somewhere in their beginning discussion.)

Your question must be clear and reasonably precise. It is generally of the form:

a. What explains Y (the dependent variable)? E.g., why is it some people vote and others do not?

But may also take the general form of

b. What is the impact of X (the independent or causative variable)? E.g. Some states have adopted term limits for their legislators. What effect does this have on the politics or policies of the state?

Or

c. Are X and Y truly related and / or is that relationship causal? E.g., We know from prior research that the state of the economy is related to a president’s popularity / election prospects. Do people really think about the economy as they cast their votes or is the relationship merely a spurious happenstance?

CHECKPOINT: Stop at this stage and verify the following about your research question

a. Is it an explanatory question (acceptable) or only a descriptive one (not acceptable)? Does it seek to explore relationships between variables?

b. Is the explanation sought general (acceptable) or ideographic (not acceptable)? Does it seek to explain a category of events or only a single event?

c. Are the important variables you are focusing on (Y, X or the relationship between the two) truly variables? Can you identify two or more exhaustive and mutually exclusive categories for each variable?

2. Why is the answer to this question important?

a. Why is it important to the scholarly community? Why is it relevant to our theoretical understanding of political institutions, behavior, or phenomena? Here an author may develop a logical argument showing how this question relates to existing theories or ideas, or may merely cite a number of prior studies of the question to demonstrate that the discipline thinks it is important.

b. Does answer the question provide practical guidance, such as resolving or understanding a contemporary policy issue?

NOTE: practical relevance can add extra value to your research BUT it is neither necessary nor sufficient. Important research is addressed to other scholars – a contribution to our understanding of the world (theory). Scholars work under the assumption of knowledge for knowledge’s sake – that understanding the world is important in and of itself and should eventually lead to the betterment of mankind even if practical applications are not immediately apparent.

3. Why is it we do not yet have a satisfactory answer to this question?

Here you conduct a brief review of past literature focused specifically on your particular research question. Why is it the question has not yet been answered to your satisfaction? Has it never been examined systematically? Do past studies disagree on their answers? Did past studies use problematic data, variable measures, or models? Does a theory produce additional theoretical implications that deserve to be examined to refine the theory or test its limits.

Why we do not have a satisfactory answeremphasizes scholarly creativity and critical thinking. You use your knowledge existing research to examine prior research theoretically: are there logical weaknesses in prior theoretical arguments or untested implications of the theory? You also use your knowledge of research design to examine prior research methodologically (broadly speaking): are there weaknesses in prior data, measures, models, strategies (e.g., experiments versus surveys), statistical techniques, necessary control variables, etc.

4. How will your proposed research bring us closer to answering the question?

Having identified some weaknesses or gaps in past research you then include a brief statement regarding how your proposed research will overcome or compensate for those weaknesses.

(NOTE: Scholars will often reverse the order of, combine, or otherwise restructure Items III and IV and sometimes also Item II below. This is solely a matter of style. You may structure Items II, III, and IV however fits best with your particular research. But all three of these items must be included and it must be clear to the reader what you are doing when you cover each.)

II. EXPANSION OF THE GAP IN THE LITERATURE:

In the Beginning Material / Introduction you cover items 3 & 4 above only briefly – enough to establish a gap in the literature and briefly state how you will help fill that gap. Here you expand on item #3 and possibly also item #2. Give a history of past research on the topic, combining your creativity with your scholarly knowledge of the literature, research design, and methodology / statistics to explain why you do not believe it answers your question. Depending upon the nature of your research and how well known the topic is in the discipline, the history can be relatively brief or much longer. In reviewing weaknesses in the prior literature discuss only problems that you expect to resolve. Do not discuss gaps or problems your research does not address unless there is some clear reason to do so. When you describe a problem with the prior literature you are implying that your research will solve this problem. Do not suggest more than you will deliver.

III. HOW TO ADDRESS THIS GAP:

What is it that should be done differently to bring us closer to answering the question. What you include here will depend upon the rationale you have selected for your research (See the document “Common Research Rationales / Justifications”).

If introducing new variables, you need to describe what they are and your reasoning as to why they should be included. At this point you may also describe how these variables will be measured / operationalized.

If arguing for different data explain what should be different from data used before and what advantages such new data would provide.

If arguing for a different model or methodological approach, explain why this will offer benefits over previous research.

HOW WILL YOU KNOW IF YOU ARE WRONG? All research is conducted because it at least has the potential to produce different results from prior research. For example, you re-examine an established model with updated data. There is no reason why such a study would be interesting unless you make the theoretical / logical argument that things might have changed across time. Similarly, you do not necessarily introduce a new variable (causative or control) into a model unless there is a logical reason to believe adding the extra variable(s) may produce results different from prior studies. It is not necessary that the results actually do change, just that there is a reason to believe they might, a possibility that needs to be examined.

A good scientific test includes the risk we might be wrong. The risk comes from stating our expectations in advance. If I am correct in my thinking what should my data show? These are always stated explicitly before you get to any analysis of the data. E.g., If my thinking (or person X’s argument) is right we expect the empirical data to show _____. Frequently these are set aside in the narrative as hypotheses. Sometimes they are not separated out as distinct hypotheses but rather included in the narrative as “expectations.” If not stated in advance, there is no risk.

At this point you should create a diagram of the model you will eventually test. You may not necessarily include this diagrammatic model in your final research paper, but it is useful for your own thinking, to verify that your work makes logical sense and to keep it clear in your own mind. If you cannot create a coherent diagram of your empirical model then this would suggest that there are gaps in your thinking about the topic.

- - - - COMPLETED RESERCH VERSUS A RESEARCH DESIGN - - - - -

Items IV, V, and VI related to a completed research project. A Research Design skips treats Item IV differently, skips Items V and VI, and includes a Bibliography instead of References in Item VII. How Items IV and VII differ between completed research and a Research Design is described below.

IV. YOUR DATA – COMPLETED RESEARCH

Here you describe the data you will use to test your hypotheses / expectations. This includes how it was gathered (from published sources, via in-person interviews, etc.), the sampling method used, the time frame of the data collection and anything else relevant to your research project. At this point you also describe how you will operationalize the variables in your model, the statistical technique you will use, etc.

IV. YOUR DATA – RESEARCH DESIGN

In II and III above you identified a gap in the research and how it ought to be addressed. For a Research Design Item IV describes what you expect to do to provide a better answer to the research question. It includes some or all of the following:

What information will you collect? Will you do your own survey, a case study, create a data file drawn from published works, or something else? Why do you expect this information will provide you an answer to your research question? How will you operationalize the variables in your model?

Techniques. How will you analyze your data? How will you decide that a relationship between (a) and (b) is confirmed or refuted by the evidence? Will you statistically assess the relationship between variables? If so, how? With an OLS model, and MLE model, or some other approach? If proposing a case study or some possible non-quantitative approach what sort of results will support or challenge your expectations?

What is the projected outline of the final written product? For a thesis or dissertation, what chapters do you expect to end up with and the basic content of each. For many this may be straightforward, a separate chapter for Items I through VII. For others you may combine some chapters or add additional ones. For example you may do both a quantitative analysis and a case study and will propose a separate chapter just for the case study.

What timetable will you have for the project? Give an estimate of the time it will take to complete each step above. When will you have Items I and II written up in a full draft form? How long will it take to complete the data collection and then the analysis? Be realistic for a dissertation committee will use this timetable to both evaluate the viability of your proposed research as well as to evaluate your progress.

In all of the above, be realistic. If this is a proposal to solicit grant funding it might be a multi-year project requiring heavy financing (your grant request). But for most projects it needs to be something doable within the time and resources available to you. For a conference paper or journal article it should be something that can be completed within about a semester – a semester when you will also be taking classes. For a dissertation proposal you will be finished with coursework and able to devote full time to the project, but it still must be something you can complete,including the full write-up of your dissertation, within about 12-18 months. You are not likely to receive department funding for more than this length of time.

V. YOUR RESULTS

RESULTS ARE NOT CRITICAL TO GOOD SCIENCE, but are important to publication potential. Too often beginning students believe that if their expectations are not borne out – none of their hypotheses confirmed, that their research is a failure. Not necessarily so. Good research is dependent upon the quality of the arguments you have made in prior sections. Is there really a gap in the literature? Is your approach a logical, theoretically and statistically sound way of addressing the gap. If so, you have conducted good research regardless of the results.

Now it is often the case, particularly for beginning students, that research fails because of sloppiness or a lack of familiarity with the literature or research techniques. Important variables might not be controlled for. Models or variable measurements are not a good match for the theory. Or the theory itself is not a complete and logically coherent argument. But if you have satisfied yourself, and your examining committee, that you do have a well-crafted research design, then null results are not a problem. Your MA thesis or dissertation will still be accepted by your committee as evidence of your research ability and thus qualifying you for a degree.

Null results, unless they disconfirm a previously accepted or proposed theory, are unlikely to be publishable. Journals emphasize research that adds to knowledge, not reports of failures to do so. But this is only a publication issue, not a research quality issue.

VI. DISCUSSION / SUGGESTIONS FOR FUTURE RESEARCH

All research is limited. We do not have perfect data, perfect measures or perfect understanding of phenomena. As you examine your results it is likely a number of other questions will arrive, questions you cannot currently address. It is common, almost standard, for scholars to end their research papers with a summary of what they have found along with a look to the future. The latter is called “suggestions for future research.” You might comment on the need for better data or measures, or some of your results might lead you to suggest additional hypotheses or tests that could be of value.

VII. REFERENCES OR BIBLIOGRAPHY

See the “Literature Review” document for a discussion of references and bibliographies and the difference between them.