A thesis statement for a class essay and one for original research aren’t quite the same thing. An essay thesis usually responds to a prompt someone else wrote. A research thesis has to survive contact with evidence you haven’t gathered yet, which means it needs to be built differently: specific enough to guide months of work, but flexible enough to survive being wrong.
This matters more than it might seem. A vague or overly broad thesis is the single most common reason an otherwise solid research project reads as unfocused. Here are 10 steps to building one that holds up.
Key takeaways
- Start from a question, not an answer. Your thesis is simply your best current answer to a question you can state in one sentence.
- A thesis and a hypothesis aren’t interchangeable. A hypothesis is a testable prediction, primarily used in quantitative work, whereas a thesis statement is your paper’s overall argument and applies to any type of research.
- Your first thesis is a working thesis, not a final one. Expect to revise it as you gather evidence; this is normal, not a sign you did something wrong.
- A thesis needs to be arguable. If a reasonable reader couldn’t disagree with it, it’s a fact or an observation, not a thesis.
- Scope kills more theses than weak writing does. A thesis broad enough to need a book gets crammed into a 15- to 20-page paper and collapses under its own size.
- Horizon mentors guide you through this process but don’t assign your thesis for you, since choosing and defending your own position is part of what makes the research yours.
1. Start with a research question, not a claim
Before you write a thesis, write the question your entire project is trying to answer. Almost any research assignment, however open-ended, can be reduced to a single question you’re trying to resolve. Your thesis is just your answer to that question once you’ve done enough work to have one.
Writing the question down first also protects you from a common trap: starting with a conclusion you already believe and then hunting for evidence that fits it. A clear question keeps you honest about what you’re actually investigating.
2. Do enough background research before you commit to a position
Don’t lock in a thesis before you know what’s already been argued or found in your area. Original research is judged partly on whether it adds something beyond what’s already established, and you can’t know what that is without reading first. This is also where you’ll discover if your question has already been thoroughly answered elsewhere, in which case you need to narrow or redirect it before going further.
3. Decide whether you need a thesis statement or a hypothesis
These aren’t the same thing, and mixing them up leads to a lot of wasted revision. A thesis statement is your paper’s central argument and applies to any type of research, quantitative, qualitative, or mixed. A hypothesis is a specific, testable prediction about the relationship between variables, and it primarily appears in quantitative or experimental work.
If your project involves measurable data, such as a survey, an experiment, or a statistical analysis, you likely need both a hypothesis predicting a specific relationship and a broader thesis that explains why that relationship matters. If your project is interpretive or literature-based, in history, literary analysis, or philosophy, for example, you need a thesis without a hypothesis.
Horizon covers both ends of this spectrum, from quantitative fields like data science and biology to qualitative ones like political theory and literature, and mentors help you figure out which structure your specific project actually calls for.
4. Write a working thesis you’re allowed to be wrong about
Your first attempt at a thesis is called a working thesis, and it’s meant to be simple and provisional. Its job is to give you a direction to start researching and writing in, not to be the polished claim that ends up in your final paper. Treating your first draft as permanent is what makes people afraid to revise it later, even when the evidence clearly points somewhere else.
5. Make sure your thesis takes a position, not just an observation or fact
A thesis has to be arguable, meaning a thoughtful reader could reasonably disagree with it and would need your evidence to be convinced. Two common failure modes to check for:
- Pure observation: “This novel touches on themes of isolation” describes something without making a claim about it.
- Statement of fact: “Many people are concerned about AI safety” is true, but nobody would need a paper to be convinced of it.
If everyone already agrees with your thesis before reading a word of your paper, it isn’t a thesis yet.
6. Narrow the scope until it’s specific enough to defend
A thesis broad enough to require a book will collapse inside a paper that’s realistically 15 to 20 pages. “Social media affects mental health” can’t be argued in any depth at that length, but “increased passive scrolling on short-form video correlates with reduced self-reported focus in adolescents” can. If you can’t picture the specific evidence you’d bring to support each part of your thesis, it’s still too broad.
7. Run it through the “so what” test
Read your thesis and ask what a skeptical reader’s first reaction would be. If the honest answer is “so what,” your thesis needs to connect to something that actually matters, either to your field, to a real-world problem, or to a gap in existing research. This is also where originality gets tested: a thesis that only restates the existing consensus in your field doesn’t clear this bar, even if it’s well argued.
8. Anticipate a counterargument or alternative explanation
Before you finalize anything, try to argue against your own thesis. If you can’t think of any reasonable objection or alternative explanation, your claim may be too obvious to be worth defending at length. If you can identify a real counterargument, addressing it directly in your paper will make your thesis considerably stronger, since it shows you’ve considered the alternative and still find your position more convincing.
9. Word it as one or two clear sentences, and place it deliberately
Keep your thesis to one or two sentences, positioned at the end of your introduction, after you’ve given the reader enough context to understand why the question matters. Avoid cramming more than one main idea into it. A useful structural template, especially for argumentative or interpretive theses: state your position, then explain why in a “because” clause, and acknowledge the strongest opposing view in an “although” clause. This keeps the sentence specific instead of vague.
10. Expect to revise it as your research develops
It’s normal and expected for your thesis to change once you’re deep into your research. If your evidence and your thesis stop lining up, the thesis is what should move, not the evidence you report. This is exactly how Horizon’s mentorship structure is built: over the course of a trimester, you revise your thesis alongside your mentor as your reading and findings develop, rather than being locked into whatever you wrote in week one.
Where does Horizon fit into this?
Horizon pairs high school students with PhD scholars and professors from Oxford, Stanford, Columbia, and other universities to produce a 20-page, university-level research paper across 600 or more specializations. When you apply, you select a general research area from a list of pre-approved topics, and once admitted, your mentor helps you refine that into a specific research proposal and thesis. Choosing your own thesis statement and revising it as your research develops remains your responsibility throughout; mentors guide the process but don’t assign the position for you, since that decision is part of what makes the research yours.
If you haven’t settled on a topic yet, this guide to choosing a research topic is a good starting point, and once your thesis and paper are further along, 10 steps to publish your research covers what comes next. You can see full program details at horizoninspires.com.
Frequently asked questions
Does every research paper need a hypothesis?
No. A hypothesis is mainly relevant to quantitative or experimental research, where you’re predicting a testable relationship between variables. Qualitative and interpretive research, in fields like history or literature, typically needs a thesis statement without a formal hypothesis.
Can your thesis change after you’ve started writing your paper?
Yes, and it often should. A thesis you write before doing the bulk of your research is a working thesis, and it’s expected to evolve as your evidence develops. Refusing to revise it when the evidence points elsewhere usually produces a weaker paper, not a more consistent one.
How long should a thesis statement be?
One to two sentences is standard. If it takes a full paragraph to state your thesis, it likely contains more than one idea and needs to be narrowed down to its central claim.
What’s the difference between a research question and a thesis statement?
A research question is what you’re investigating, phrased as a question. A thesis statement is your answer to that question, phrased as a claim, once you’ve done enough research to defend it.
Can two students research the same general topic with different thesis statements?
Yes, and this is common. The same broad topic can support many different research questions, and two students investigating similar territory will often land on entirely different theses depending on which angle, population, or time period they focus on.
What if your mentor or teacher disagrees with your thesis?
Treat it as a stress test, not a verdict. Ask specifically what part they find unconvincing, whether it’s the scope, the evidence, or the arguability, since that tells you what to revise. Disagreement at this stage is far cheaper to address than discovering the same problem after months of research.
Image source: Horizon Academic Research Program