A research question in psychology is a focused, answerable question that identifies a specific gap in knowledge about human behavior, cognition, or emotion and guides every decision in a study’s design. It’s the starting point of the scientific process: before a psychologist chooses participants, picks a method, or runs a single statistical test, they need a clear question that tells them what they’re looking for and why it matters.
If you’re encountering this concept for the first time, you’re likely planning a research project or trying to make sense of published studies. Either way, understanding how research questions work will sharpen both your writing and your ability to evaluate psychological science.
What a Research Question Actually Does
A research question serves as the backbone of a study. It identifies the problem being investigated, points toward the right methodology, and sets the stage for building a testable hypothesis. Without one, a study has no focus. With a vague one, the results are difficult to interpret.
Think of it as a filter. Psychology is enormous, spanning everything from childhood development to decision-making under stress to the effects of social media on mood. A research question takes that broad territory and narrows it to something a single study can meaningfully address. “How does social media affect people?” is a topic. “Do adolescents who use social media for more than three hours daily report higher levels of anxiety than those who use it for less than one hour?” is a research question. The difference is precision.
Research Questions vs. Hypotheses
These two terms get confused constantly, but they play different roles. A research question asks what you want to find out. A hypothesis is a specific, testable prediction about what you expect the answer to be. The question comes first, and the hypothesis grows out of it.
For example, your research question might be: “Does sleep deprivation affect working memory in college students?” Your hypothesis would then predict a direction: “Students who sleep fewer than five hours will perform worse on a working memory task than students who sleep seven or more hours.” When it comes time to run statistics, the hypothesis gets restated as a “null hypothesis,” which assumes there’s no difference between groups. The entire analysis is built around trying to reject that null assumption. But none of that statistical machinery works without a clear research question driving it from the start.
Three Main Types
Research questions in psychology generally fall into three categories, and the type you choose determines the kind of study you’ll design.
- Descriptive questions ask what something looks like. They document patterns, frequencies, or characteristics without trying to explain why. Example: “What percentage of first-year university students report symptoms of test anxiety?” These questions often use surveys or observational methods.
- Relational (predictive) questions ask whether two things are connected. They look for associations or correlations but don’t claim one thing causes the other. Example: “Is there a relationship between hours of weekly exercise and self-reported stress levels among working adults?” These questions typically use correlational designs.
- Causal questions ask whether one thing directly causes a change in another. They require experimental designs where the researcher controls variables. Example: “Does cognitive behavioral therapy reduce panic attack frequency more than a waitlist control over 12 weeks?” Only experiments with proper controls can answer causal questions.
Knowing which type you’re working with matters because it determines what conclusions you’re allowed to draw. A correlational study can’t answer a causal question, no matter how strong the association looks.
How to Build a Strong Research Question
The most common mistake students make is starting too broad. “What causes depression?” isn’t a research question. It’s a career. Narrowing down requires you to specify who you’re studying, what you’re measuring, and under what conditions.
One widely used framework for this is PICOT, which stands for Population, Intervention, Comparison, Outcome, and Time. You identify the group of people you want to study, the variable or treatment you’re interested in, what you’re comparing it against, the outcome you’ll measure, and the timeframe for data collection. Not every research question uses all five elements, but running through them forces you to get specific. A question like “Does mindfulness training reduce test anxiety in undergraduate students compared to a study-skills workshop over one semester?” hits every PICOT component.
The other critical step is operationalization, which means defining your psychological concepts in terms that can actually be measured. “Anxiety” is too abstract to study directly. You need to decide whether you’re measuring it through a validated questionnaire, physiological markers like heart rate, behavioral observations, or something else entirely. Two studies can ask seemingly identical questions but produce very different results depending on how they operationalize their variables. Making those choices explicit in your research question (or immediately after it) keeps the entire project grounded.
What Makes a Research Question Good
A useful evaluation tool is the FINER criteria, which checks five qualities:
- Feasible: Can you realistically answer this question with the time, funding, skills, and data you have access to? A brilliant question you can’t actually investigate isn’t useful.
- Interesting: Does the question matter to you and to the broader scientific community? If no one cares about the answer, the study won’t contribute much.
- Novel: Does it address a genuine gap in existing knowledge? A thorough literature review will reveal what’s already been studied and where the open questions remain.
- Ethical: Can the study be conducted without causing harm to participants? Research involving human subjects must go through an institutional review board (IRB), which evaluates whether risks are minimized, informed consent is obtained, and vulnerable populations are adequately protected.
- Relevant: Will the findings have practical value, whether for clinical practice, education, policy, or simply deeper understanding of human psychology?
The ethical dimension deserves extra emphasis in psychology. Some questions that would be fascinating to answer simply can’t be studied in the way you’d want. You can’t randomly assign children to neglectful environments to study the effects of neglect. You can’t withhold effective treatment from people in crisis to maintain a control group. The IRB exists to catch these problems before a study begins, and your research question needs to be designed with those boundaries in mind from the start.
Examples Across Psychology Subfields
Research questions look different depending on which area of psychology you’re working in, and seeing concrete examples helps clarify what a well-formed question looks like in practice.
In clinical psychology, questions often focus on treatment effectiveness or the nature of mental health conditions: “Do clients report feeling safer in online therapy sessions compared to in-person sessions?” or “Does trauma-focused somatic therapy produce lasting symptom reduction comparable to cognitive behavioral therapy?” These questions are timely, specific, and testable.
In social psychology, the focus shifts to how people influence each other: “Does exposure to racially diverse media reduce implicit bias scores on a standardized measure?” In cognitive psychology, questions target mental processes: “Does bilingualism slow age-related decline in task-switching ability?” In developmental psychology, you might ask: “At what age do children begin to understand that others can hold false beliefs, and does this vary across cultures?”
Each of these examples identifies a population, specifies what’s being measured, and implies a research design. That’s the hallmark of a question that’s ready to guide an actual study rather than just spark a conversation.
Common Pitfalls to Avoid
Questions that are too broad (“How does stress affect health?”) give you no direction for designing a study. Questions that are too narrow (“Do left-handed 19-year-old women in rural Ohio sleep differently on Tuesdays?”) produce findings no one can generalize. The sweet spot is specific enough to be answerable but broad enough to be meaningful.
Another frequent problem is embedding assumptions into the question. “Why does social media cause depression in teenagers?” assumes a causal link that hasn’t been established. A better version would ask whether an association exists before jumping to causation. Similarly, avoid questions that can be answered with a simple yes or no. “Does therapy work?” doesn’t tell you what kind of therapy, for whom, compared to what, or measured how. Every vague word in a research question is a decision you haven’t made yet, and those unmade decisions will haunt you when it’s time to collect and interpret data.
Finally, make sure your question is actually answerable with empirical evidence. “Is it morally wrong to lie?” is a philosophical question. “Do people who frequently lie report lower life satisfaction?” is a psychological one. The distinction is whether data can resolve it.