If you're not sure whether to write a hypothesis or a research question, you're probably doing quantitative vs qualitative — and your instructor hasn't spelled out which lane you're in. The distinction matters more than it looks. A research question asks something open; a hypothesis predicts a specific, testable relationship between variables. Using the wrong one in your introduction confuses reviewers and weakens your Methods section, because the analysis plan for each is different.
PaperDraft is a writing assistant, not a paper generator — the draft is your starting point, not your submission. You are responsible for editing, verifying sources, and following your school's academic integrity policy.
This guide explains when to use a research question vs hypothesis, how to write each one, and the most common errors students make. For the full paper arc, see how to write a research paper.
What a Research Question Actually Is
A research question is a specific, focused inquiry your study is designed to answer. It doesn't predict an outcome — it asks.
Good research questions share four traits:
- Specific. Names the variables, population, and context.
- Answerable. Can be addressed with the data you plan to collect.
- Scoped. Narrow enough to cover in one paper.
- Grounded. Emerges from a gap you identified in the literature.
Examples of solid research questions:
- "How do first-generation undergraduates describe their experience of academic belonging in their first semester at large public universities?"
- "What themes emerge in nurses' narratives of moral distress during end-of-life care?"
- "How did immigrant textile workers in 1910s Massachusetts describe their working conditions in Yiddish-language newspapers?"
Notice: each is open-ended. The answer requires analysis, not just a yes/no result.
What a Hypothesis Actually Is
A hypothesis is a testable prediction about the relationship between variables. It names the variables, the direction of the relationship, and (in quantitative work) the population.
Good hypotheses share four traits:
- Testable. You can collect data that would confirm or disconfirm it.
- Specific. Names the independent and dependent variables.
- Falsifiable. There's a result that would clearly reject it.
- Grounded in theory. The prediction comes from a framework or prior evidence, not a guess.
Examples of solid hypotheses:
- "Undergraduates in the mindfulness condition will report lower state anxiety than those in the control condition."
- "Patients in the family-presence protocol will have lower rates of postoperative delirium than those in standard care."
- "Higher perceived team autonomy will predict lower emotional exhaustion in hybrid knowledge workers, controlling for workload."
Each one predicts a direction and is falsifiable — the data can reject it.
When to Use Which
The decision isn't a preference. It follows your study design.
Use a research question when
- Your study is qualitative — interviews, focus groups, ethnography, narrative analysis, archival work.
- You're doing exploratory work where no strong theory predicts a direction.
- You're doing a mixed-methods study and need an umbrella inquiry that spans both parts.
- Your discipline (much of the humanities, anthropology, some education research) frames inquiry as questions rather than predictions.
Use a hypothesis when
- Your study is quantitative — experiments, surveys with statistical testing, longitudinal data.
- Theory or prior research predicts a specific direction.
- Your Methods section will include hypothesis-testing statistics (t-tests, ANOVAs, regression).
Use both when
- Your study is mixed-methods and the quantitative arm tests hypotheses while the qualitative arm explores questions.
- Some fields (education, health sciences) ask for an overarching research question and specific hypotheses that operationalize it.
Check your assignment. If it says "state your hypothesis" and you're doing an interview study, clarify with your instructor before forcing the wrong fit.
How to Write a Good Research Question (Template and Examples)
A working template for qualitative questions:
"How do [population] [experience / describe / make sense of] [phenomenon] in [context]?"
Examples:
- "How do graduate students describe their experience of imposter syndrome during dissertation writing?"
- "How do rural primary care physicians make sense of telehealth adoption in their practices?"
- "What meanings do high school teachers attach to teaching controversial historical events?"
Avoid yes/no framings ("Do students experience imposter syndrome?"). Open with "How," "What," or "In what ways."
How to Write a Good Hypothesis (Template and Examples)
A working template for quantitative hypotheses:
"[Independent variable] will [increase / decrease / predict] [dependent variable] in [population], [controlling for X]."
Examples:
- "A gratitude journaling intervention will reduce depressive symptoms in undergraduates over a four-week period."
- "Sleep deprivation will impair working memory performance on the n-back task in adults aged 18-25."
- "Autonomy-supportive teaching will predict higher intrinsic motivation in asynchronous online learners, controlling for prior GPA."
Keep direction explicit. A directional hypothesis is stronger than a non-directional one when theory justifies the direction.
Null Hypothesis, Alternative Hypothesis, and One-Tailed vs Two-Tailed
If your Methods section uses inferential statistics, you'll also need a null hypothesis (H0) and alternative hypothesis (H1). The null states there is no effect; the alternative states there is.
- H0: There is no difference in anxiety scores between the mindfulness and control conditions.
- H1: There is a difference in anxiety scores between the mindfulness and control conditions.
If you predict a direction, the H1 is directional. If you don't, it's two-tailed. For worked examples across fields, see null hypothesis examples.
Clear on whether it's a question or hypothesis, but still staring at a blank Introduction? PaperDraft gives you a structured first draft — thesis stub, IMRaD skeleton, opening sections in academic register — so you can spend your time refining your prediction instead of formatting. It's a drafting assistant, not a submission. Try PaperDraft — free
Common Mistakes Students Make
A few errors show up repeatedly, regardless of field.
Writing a hypothesis for a qualitative study. "I hypothesize that first-generation students will feel isolated." You can't falsify that with interview data. Reframe as a research question.
Writing a research question for an experimental study. "Does mindfulness reduce anxiety?" is technically a question, but in a randomized trial it should be a directional hypothesis.
Non-specific hypotheses. "Social media will affect well-being" is too vague. Which platform? Which measure of well-being? Which population?
Unfalsifiable hypotheses. "The intervention will have some effect on some outcome" can't be rejected. Make the prediction sharp enough to fail.
Mismatching hypothesis and analysis. You hypothesize an interaction, but your Results section only reports main effects. The two have to align.
Too many hypotheses. A 15-page paper with 11 hypotheses is a Results buffet. Two to four is typical.
How a Drafting Assistant Fits
A drafting tool can sketch question and hypothesis templates in the correct register, scaffold your introduction around them, and flag when the framing doesn't match the Methods section you've described. What it can't do is decide whether your study is quantitative or qualitative — that's a design choice you've made — or validate that your hypothesis is grounded in the right theory for your field. PaperDraft handles the structure and language. You handle the design logic and the literature grounding.
FAQ
Can a paper have both a research question and a hypothesis?
Yes, especially in mixed-methods designs or when a broad research question is operationalized by specific hypotheses.
Where does the hypothesis go in the paper?
At the end of the Introduction, right before the Methods section. The question or hypothesis is the last thing the reader sees before the study design.
Do I need to state the null hypothesis explicitly?
In most write-ups, no — the null is implied when you test an alternative. Some fields (statistics-heavy, some health sciences) ask for both explicitly.
What if my data don't support my hypothesis?
Report it honestly. Non-significant results are still results. The Discussion explains what happened and what the null result means for theory.
Can a research question evolve during the study?
In qualitative work, yes. Grounded theory explicitly allows questions to refine as data come in. In quantitative work, changing your hypothesis after seeing results is HARKing — a research-integrity problem. Pre-register if you can.
Once your question or hypothesis is sharp, the rest of the paper has a spine to hang on. For the next step — designing the study that answers it — see sample size in research.