Steps in the Research Process: A Guide for Students

The research process is the series of steps you follow to answer a question with evidence. You pick a topic, turn it into a research question, read what is already known, choose a method, collect and analyze data, and write up what you found. Every college research paper, lab project and capstone follows some version of it, even when the assignment sheet only says "write a 10-page paper."

This guide walks through seven steps using one running example: a student project on phone use before bed and sleep among college students. At each stage you'll see the actual question, search string, source notes, survey items and write-up plan, plus tables you can adapt.

What Is the Research Process?

The research process is a structured way of moving from curiosity to an answer you can defend. It's what separates a research paper from an opinion essay: every claim traces back to a source or to data you collected, and a reader can see how you got there.

It applies to a survey for a psychology methods course and to a library paper for English composition alike. The tools change. The logic doesn't.

Why Guides List Different Numbers of Steps

Guides list anywhere from five to eleven steps because they split the same work in different places. John Creswell's widely used textbook Educational Research groups it into six: identify a research problem, review the literature, specify a purpose, collect data, analyze and interpret the data, and report and evaluate the research.

Cornell University's well-known library guide, "The Seven Steps of the Research Process," adapted in this open writing textbook's chapter on the seven steps of research, focuses on library work instead: develop a topic, find background information, find books, find articles, find internet resources, evaluate what you find and cite it.

Neither is wrong. A lab course stresses design and data collection; a history course stresses finding and evaluating sources. The seven steps below cover both.

The Process Loops Back on Itself

Research rarely runs in a straight line. The Association of College and Research Libraries, whose Framework for Information Literacy for Higher Education guides research instruction at many colleges, describes research as inquiry: it is iterative, and answering one question usually leads to new, more complex questions.

In practice, your reading may show that your question has already been answered, or that it's too broad to finish in six weeks. When that happens, you go back a step and revise. That's normal.

The Running Example Used in This Guide

Imagine your research methods course asks for a small original study, and you're curious about phones and sleep because you and your roommates scroll in bed every night. That project is the example used in every step below.

The 7 Steps in the Research Process at a Glance

The seven steps in the research process are: choose a topic, write a research question, review the literature, choose a research design, collect data, analyze and interpret it, and write up your findings. The table shows what each step should produce and how it looks in the running example.

How the Steps Change for a Source-Based Paper

If your assignment is a library research paper rather than an original study, your sources are your data. Steps 1 to 3 stay the same. Step 4 becomes deciding which sources count (peer-reviewed studies only? the last ten years?) and how you'll compare them. Step 5 becomes a systematic search with careful notes. Step 6 becomes synthesis: grouping sources by what they found, where they agree and where they conflict. Your thesis comes out of that synthesis, not before it.

Step 1: Choose a Topic and Narrow It

Start with a topic you care about, then cut it down until you could realistically finish it with the time, sources and access you have. "Technology and health" is a subject area, not a research topic. Nobody could cover it in a semester.

Narrow it by adding limits one at a time:

• Who: college students, not "people"
• What: phone use, not "technology"
• When: the hour before bed, not "all day"
• Outcome: sleep, not "health"
• Where: undergraduates at your college

In the running example, the topic moves from "technology and health" to "social media and sleep" to "phone use in the hour before bed and weeknight sleep among undergraduates at one college." That last version tells you who to study, what to measure and where to find participants.

Before you commit, check feasibility. Can you find a handful of scholarly sources? Can you reach the people or data you need? Does it fit the deadline? If you're still looking for a direction, browse these research paper topics and tips for choosing one.

Common mistake: picking a topic so new or so local that there's nothing to read about it, or so famous that you can't say anything fresh.

Step 2: Turn Your Topic into a Research Question

Your research question is the one question your whole project exists to answer, so write it before you collect anything. A good one is focused, answerable with evidence you can actually get, and open enough that the answer isn't obvious. A review in the Canadian Journal of Surgery on writing research questions, hypotheses and objectives calls developing the research question "the most important aspect of a research project."

Weak vs. Strong Research Questions

Notice the word "associated." A one-time survey can show that two things go together. It can't show that one causes the other. Matching your wording to what your design can prove saves you from overclaiming later. For more examples across subjects, see what makes a good research question.

Check It Against the FINER Criteria

The FINER criteria, which that same review credits to Hulley and colleagues' textbook Designing Clinical Research, give you a quick five-point test. Here's the running example's question run through it:

• Feasible: enough participants, skills, time and money, with a manageable scope? Yes. An online survey of classmates is free and takes a few weeks.
• Interesting: would the answer interest you, your peers and your instructor? Most students have an opinion about phones and sleep.
• Novel: does it confirm, refute or extend earlier findings? It tests whether patterns reported in younger students show up on your campus.
• Ethical: would an ethics review board approve it? An anonymous, voluntary survey of adults is low risk.
• Relevant: does it matter for knowledge, policy or future research? A campus wellness office could use the answer.

Add a Hypothesis If Your Design Needs One

A hypothesis is your predicted answer, stated so the data can support or contradict it. Quantitative studies usually need one; exploratory qualitative studies usually don't. For the running example: "Students who report more than 30 minutes of phone use in the hour before bed will report fewer hours of weeknight sleep than students who report 30 minutes or less." Write it down now, before you see any data, so you can't quietly bend it to fit your results later.

Step 3: Conduct a Literature Review

A literature review is a structured search and synthesis of what researchers already know about your question. It tells you whether your question has been answered, which methods others used and where your project fits. Even on a small class project, it's where you find your measures, comparison points and most of your citations.

Build a Search String

Begin with background reading so you learn the field's vocabulary. For sleep, the CDC's About Sleep page with recommended hours by age says adults ages 18 to 60 need 7 or more hours per night, which gives you a benchmark to compare your own results against.

Next, list the key ideas in your question, add synonyms for each and combine them. OR widens a search, AND narrows it, and quotation marks keep a phrase together. For the running example:

("smartphone" OR "mobile phone" OR "screen time") AND (bedtime OR "before sleep") AND ("sleep duration" OR "sleep quality") AND ("college students" OR undergraduates)

Run it in your library's databases (PsycINFO and PubMed suit this topic) and in Google Scholar. The Google Scholar search help page points out two time-savers: "Cited by" shows newer papers that referenced a result, and "Related articles" finds similar work. Links labeled with your library's name can also appear beside results your school subscribes to.

Evaluate Each Source

Not every source that matches your search belongs in your paper. The CRAAP test from Meriam Library at California State University, Chico gives five checks: Currency, Relevance, Authority, Accuracy and Purpose. For research papers, add two questions of your own: who was studied, and what design was used?

Here's how that works on a real source you'd find for the running example. Carter and colleagues' 2016 systematic review and meta-analysis in JAMA Pediatrics, "Association Between Portable Screen-Based Media Device Access or Use and Sleep Outcomes," pooled 20 studies covering 125,198 children aged 6 to 19. It found that bedtime use of a media device was significantly associated with inadequate sleep quantity, poor sleep quality and excessive daytime sleepiness.

That's a strong, credible source. But look closer. Everyone in it was 19 or younger, with a mean age of 14.5, and all 20 included studies were cross-sectional. So it supports the idea that bedtime device use and poor sleep go together in school-age children. It doesn't tell you what happens among college students, and it doesn't prove that phones cause poor sleep. Both points belong in your review. Spotting them is what instructors mean by "critical evaluation."

Synthesize Sources in a Matrix

A literature review compares sources instead of summarizing them one at a time. A source matrix makes that easier: give each source a row and each feature a column, then read down the columns to spot patterns, agreements and gaps.

If your instructor wants sources written up before the review itself, an annotated bibliography in APA or MLA format makes a natural first draft of this matrix.

Step 4: Choose Your Research Design

Your research design is the plan for how you'll answer the question: what kind of data you'll collect, from whom and how. Let the question choose the design, not the other way round.

Quantitative, Qualitative or Mixed Methods

The running example asks whether two measurable things are related, so it calls for a quantitative design: a cross-sectional survey, where you measure everyone once. An experiment that has some students put their phones away for two weeks could test cause and effect, but it's far harder to run on a class timeline. For more on studies built on data you gather yourself, see our student guide to empirical research.

Pick Your Sample

Your population is everyone your question is about, here all full-time undergraduates at your college. Your sample is the people you actually reach. On a class project you'll usually use a convenience sample, such as students in a few large lecture courses. That's acceptable if you say so, but it limits how far you can generalize. Decide in advance who qualifies (for example, full-time undergraduates aged 18 or older) and write it down.

Check Ethics and IRB Rules Before You Start

If your project collects information from people, find out whether it needs review by your school's Institutional Review Board (IRB) before you contact anyone. The federal Common Rule, in 45 CFR 46.102, defines research as "a systematic investigation, including research development, testing, and evaluation, designed to develop or contribute to generalizable knowledge." Its definition of a human subject covers studies run by an investigator "whether professional or student."

Many colleges treat a class project as outside full IRB review when its main purpose is teaching research skills. Northeastern University's policy on classroom research is a typical example: a project can qualify when it's minimal risk, excludes people under 18 and other vulnerable groups, avoids sensitive information and isn't expected to be published or presented publicly. Theses and dissertations still go to the IRB. Rules differ by school, so ask your instructor first.

For the running example, that means an anonymous survey, a short consent statement on the first screen, a "prefer not to answer" option on every item and no names or student ID numbers.

Step 5: Plan and Carry Out Data Collection

Data collection is the step where you gather your evidence, and the rule is consistency: collect it the same way from everyone, exactly as your method plan says. Changing questions halfway through makes responses impossible to compare.

Write Questions That Measure One Thing

Each survey item should measure one variable. The Pew Research Center's guide to writing survey questions warns against double-barreled questions, which ask people to evaluate more than one concept at once, and notes that "even small wording differences can substantially affect the answers people provide."

Where you can, borrow a validated measure instead of writing your own. For sleep, a well-known one is the Pittsburgh Sleep Quality Index (PSQI), a self-rated questionnaire whose 19 items produce seven component scores and one global score covering the past month. Ask your instructor whether to use it, and check whether you need permission.

Pilot Test Before You Launch

Pew calls pretesting "an essential step in the questionnaire design process." Send your survey to three to five people who fit your sample, time how long it takes and ask which questions confused them. Fix the problems before the real launch, and keep pilot responses out of your final data.

While you collect, keep a research log: when the survey opened and closed, where you shared the link, how many people responded and every decision you made. Set cleaning rules (for example, dropping responses that skip both sleep questions) before you look at results. The log becomes the backbone of your methods section.

Step 6: Analyze and Interpret Your Data

Analysis turns raw responses into an answer to your research question, and interpretation explains what that answer means and how far it reaches. Do them in that order.

Describe, Then Test

Start with descriptive statistics: how many people responded, who they were, and the average and spread of each variable. For the running example, you'd report median weeknight sleep and the share of students sleeping under 7 hours, the CDC benchmark from your literature review.

Then compare. Split respondents into two groups, 30 minutes or less of bedtime phone use versus more than 30 minutes, and compare their average sleep. Because you collected minutes as numbers, you can also correlate phone minutes with sleep hours. If your course covers inferential statistics, a t-test or correlation test tells you whether the pattern is larger than chance alone would likely produce. If you're unsure which applies, review the difference between descriptive and inferential statistics.

A results sentence should give the numbers and nothing more. Its shape looks like this, with placeholders instead of real findings: "Students who reported more than 30 minutes of bedtime phone use slept a mean of X hours on weeknights (n = __), compared with Y hours among students who reported 30 minutes or less (n = __)."

Say What Your Results Can and Cannot Show

Interpretation is where many student papers overreach. Even if heavy bedtime phone users in your sample sleep less, a cross-sectional survey can't show that phones caused it. Other explanations fit the same pattern: students with jobs, heavy course loads or high stress might both stay up on their phones and sleep less. Name these alternatives yourself before your reader does.

Then compare your findings with your literature review. Do they match what Carter and colleagues found in younger students, or not? Finally, list your limitations plainly: self-reported sleep, a convenience sample from one campus and a single point in time. Limitations don't weaken a paper. Hiding them does.

Common mistake: writing "proves" when you mean "suggests," or treating a result from a few dozen classmates as true of all college students.

Step 7: Write Up, Cite and Revise

The last of the steps in the research process is reporting what you did and found, clearly enough that a reader could check your work or repeat it. Draft from your matrix, log and results tables rather than from memory.

Use the IMRaD Structure

Most empirical papers follow IMRaD: Introduction, Methods, Results and Discussion. A study in the Journal of the Medical Library Association traced how the IMRaD structure became standard and found that during the 1980s it became the only pattern used for original papers in the four major medical journals it examined.

A source-based paper usually replaces Methods and Results with thematic body sections, but it still needs a clear question up front and a direct answer at the end. Our guide on how to write a research paper outline helps you plan the sections before you draft.

Cite as You Go

Add citations while you draft, not at the end, and use the style your instructor assigns. In APA 7, the meta-analysis from the running example looks like this in the reference list:

Carter, B., Rees, P., Hale, L., Bhattacharjee, D., & Paradkar, M. S. (2016). Association between portable screen-based media device access or use and sleep outcomes: A systematic review and meta-analysis. JAMA Pediatrics, 170(12), 1202–1208. https://doi.org/10.1001/jamapediatrics.2016.2341

For in-text citations and other reference types, keep our APA style guide with 7th edition examples open while you write.

Revise Against Your Research Question

Before you submit, reread your research question and then your conclusion. The conclusion should answer the question you actually asked, at the strength your design allows. Then check that every factual claim has a citation, every table is mentioned in the text, and your methods section would let a classmate repeat the study.

If you get stuck at any step, from narrowing a topic to interpreting a correlation, you can work through your own draft with a tutor who explains the method and gives feedback, so the finished paper is still yours.

Frequently Asked Questions

What Are the 7 Steps in the Research Process?

The seven steps are: choose and narrow a topic, write a research question, conduct a literature review, choose a research design, collect data, analyze and interpret the data, and write up and cite your findings. Some guides merge or split these, so you may see five, six or eight steps. The order matters more than the count, and you'll often loop back to an earlier step when your reading or data surprise you.

What Is the First Step in the Research Process?

The first step is choosing a topic and narrowing it until you can finish it with the time and sources you have. Many guides fold the research question into this step, because a topic only becomes researchable once you can phrase it as one focused question. Background reading at this stage shows you whether the topic is too broad, too narrow or already well answered.

What Is the Most Important Step in the Research Process?

Most researchers would point to the research question. A review in the Canadian Journal of Surgery calls developing it "the most important aspect of a research project," because the question decides what you read, which design you choose, what data you collect and what your conclusion can claim. A vague question causes problems at every later step, so get feedback on it early.

What Is the Difference Between a Research Topic and a Research Question?

A topic is the subject area you're studying, such as phone use and sleep. A research question is the specific, answerable question you'll investigate within it, such as whether bedtime phone use is associated with weeknight sleep among undergraduates at your college. One topic can produce dozens of questions. Your paper needs one main question, sometimes with a few sub-questions, so every section has a clear job.

Do Students Need IRB Approval for a Class Survey?

Sometimes. Many colleges don't require IRB review for class projects whose main purpose is teaching research skills, as long as they're minimal risk, avoid minors and sensitive topics, and won't be published or presented publicly. Theses, dissertations and projects you plan to share beyond class usually need review. Federal rules explicitly cover student investigators, so ask your instructor or your school's IRB office before you contact any participants.

How Long Does the Research Process Take?

It depends on the project. A short source-based paper can move through every step in two or three weeks. An original study for a methods course usually needs five or six, because ethics checks, pilot testing and data collection all take real time, and a thesis takes months. Whatever the length, give the topic, question and literature review as much time as the writing.

What Is the Final Step in the Research Process?

The final step is reporting your findings: writing up what you asked, how you studied it, what you found and what it means, with full citations. Many models add evaluation or reflection at the end, where you note the study's limitations and the new questions it raised. Those questions often start your next project, which is why the process is often drawn as a cycle.