Research begins with a question, not an answer. A strong paper identifies something worth investigating, considers competing explanations, evaluates evidence, and reaches conclusions that the evidence can support.
To illustrate the process, this guide follows one example: Why are some countries more prosperous than others? This is one of the project ideas on my Polygence mentor profile . It offers an excellent starting point precisely because one student paper cannot answer such a broad question. The challenge is to turn it into a manageable research project.
I present the stages in order for clarity, but researchers often move back and forth between them. What you learn from sources or data may require you to rethink your question, methods, or expectations.
Develop a good research question
A topic identifies an area of interest. A research question specifies what you want to understand or explain. A useful question is clear, analytically interesting, narrow enough to investigate, and answerable with evidence you can realistically obtain.
Topic: Economic development.
Initial question: Why are some countries richer than others?
More focused question: To what extent does educational attainment explain differences in income per person across countries?
Alternative question: How do differences in political institutions influence economic development in two otherwise comparable countries?
The two focused questions imply different projects. The first invites quantitative comparison; the second may be better suited to comparative historical analysis. Neither is automatically superior. The appropriate choice depends on what you want to explain and the evidence available.
Four tests of a research question
- Clarity: Can a reader tell exactly what you are investigating?
- Analytical value: Does the question call for explanation, comparison, or evaluation rather than mere description?
- Feasibility: Can you find relevant evidence and complete the analysis in the time available?
- Scope: Have you limited the cases, time period, or relationship enough to examine it carefully?
Try it: Write your question in one sentence. Then identify the outcome you want to explain, the explanations you want to consider, and the evidence you might need.
Explore existing research
Before settling on a final question, find out what researchers have already learned. A preliminary literature review helps identify plausible explanations, disagreements, measurement problems, and available data. It may also show that your original question is too broad or rests on an assumption that needs testing.
Why do countries differ in prosperity? Researchers have examined education and skills, institutions, geography, investment, technology, trade, and historical circumstances. These explanations may reinforce one another; they are not necessarily mutually exclusive.
A student focusing on education should therefore ask whether differences in education are an independent explanation of income differences, a consequence of prosperity, or both.
Read sources analytically
For each useful study, record: its question, central explanation, type of evidence, principal finding, and limitations. Compare studies rather than compiling isolated summaries.
Try it: Find three relevant academic studies. Write two sentences explaining where they agree or disagree and what remains uncertain.
Write a research proposal
A proposal is a concise plan for answering your question. It should persuade a reader that the project is interesting, grounded in existing knowledge, and feasible. It should not assume the answer in advance.
Question: To what extent do secondary-school completion rates explain variation in income per person across countries?
Motivation: Researchers often identify education as a major source of economic prosperity, but institutions, investment, and reverse causation may also shape the relationship.
Approach: Compare public country-level education and income indicators for a recent year, visualize the relationship, and examine how regional or income groupings affect the pattern.
Limitation: A cross-country association alone cannot establish that education causes higher income.
A useful one-page structure
- Question: State the precise issue you will investigate.
- Why it matters: Explain its significance.
- Existing explanations: Summarize the most relevant research.
- Expectations: Identify hypotheses or competing arguments.
- Research design: Specify cases, measures, sources, and method.
- Feasibility and limitations: Explain what the project can and cannot establish.
Try it: Draft a one-page proposal before collecting extensive data or writing the full paper. Revise it as you learn more.
Choose a research method
Methods are tools for answering questions, not badges of sophistication. A simple comparison that answers the question well offers more value than an elaborate statistical analysis that misses the question.
| Approach | Best suited to | Possible application |
|---|---|---|
| Quantitative | Patterns and relationships across many observations | Compare education and income measures across countries |
| Qualitative | Processes, decisions, institutions, and mechanisms | Study how and why changes in education outcomes affected per capita income in one country |
| Comparative | Explaining similarities or differences between cases | Compare two countries with different development trajectories |
Start with the evidence you can obtain
For a quantitative project, check whether researchers measure variables consistently and whether the dataset contains enough observations for a meaningful comparison. For a qualitative project, consider whether documents, interviews, or historical accounts illuminate the proposed mechanism. For comparative research, explain why your chosen cases provide an informative comparison.
Distinguish association from causation
Suppose countries with higher educational attainment tend to have higher incomes. Education may raise productivity, but richer countries can also afford better schools. Other factors may influence both. Your paper can describe a pattern without claiming to have established a causal effect.
Can you identify the unit of analysis, time period, variables or cases, sources, and main limitations? If not, the question may need further narrowing.
Find and assess high-quality sources
A source is useful when it is credible and appropriate for the claim you want to make. A respected institution can publish a dataset that is poorly suited to your question; an authoritative article can still have methodological limitations.
| Source | Questions to ask |
|---|---|
| Peer-reviewed research | Which method did the researchers use? What does the evidence actually establish? |
| Official statistics | How does the source define its indicators? Can you compare countries and years? |
| Primary documents | Who wrote the document, why, and in what context? |
| Journalism and analysis | Does the author attribute and corroborate claims and distinguish them from opinion? |
For the development example, international datasets from the World Bank or UNESCO may be useful starting points. Read the indicator definitions and documentation before comparing numbers.
Write the research paper
Organize the paper around the question and the logic of your argument, not around the order in which you happened to find sources. A typical empirical paper may include an introduction, existing research, research design, analysis, discussion, and conclusion.
Each substantive section should advance the answer. Distinguish what the data show from how you interpret them, and explain why alternative explanations do or do not fit the evidence.
“Countries with higher secondary-school completion rates tend to have higher income per person in this sample. This pattern is consistent with the importance of education, but it does not by itself establish the direction or size of a causal effect.”
Revise the argument, not just the prose
Revision starts by testing the paper's reasoning. Does it answer the question it set out to investigate? Does each claim follow from the evidence? Have you considered rival explanations and acknowledged limitations?
- Make the research question and main conclusion consistent.
- Remove claims that exceed the evidence.
- Explain contradictory findings instead of ignoring them.
- Check that every table, figure, and quotation serves the argument.
- Then revise organization, clarity, citations, and style.
Strong papers do not pretend to eliminate uncertainty. They explain what researchers have learned and identify what questions remain unresolved.
Develop your own research project
I help students turn broad interests into focused research questions, design feasible studies, evaluate evidence, and develop clear, well-supported arguments. My background spans economics, political science, public policy, and quantitative research.
Barak Hoffman, Ph.D. · Former Federal Reserve and U.S. Treasury economist · University teaching and research experience
Inquire about research mentoring