What Is the Scientific Method? Definition, Steps, and Examples

The scientific method is a step-by-step process for testing ideas about the world against evidence. A researcher observes something, asks a question, proposes a testable hypothesis, and runs an experiment. The results are analyzed, a conclusion is drawn, and the findings are shared so others can check them. This guide explains each step in order, shows how the 5-, 6-, and 7-step versions line up, and works through two examples.

For where the scientific method fits among other research approaches, see the research methodology guide for graduate students.

The 7 Steps of the Scientific Method at a Glance

The scientific method is usually taught as seven steps, carried out in this order:

1. Observation: notice something specific in the world that prompts curiosity.
2. Question: turn the observation into a focused question that evidence can answer.
3. Hypothesis: propose a testable, falsifiable explanation that predicts an outcome.
4. Experiment: test the hypothesis by changing one factor and measuring the result.
5. Data analysis: analyze the measurements to see whether they fit the prediction.
6. Conclusion: decide whether the data support the hypothesis, and within what limits.
7. Communication and replication: share the findings so other researchers can check and repeat them.

What Is the Scientific Method?

The scientific method is a structured way of testing ideas so that conclusions rest on evidence, not on who is making the claim. It combines careful observation, testable predictions, measurement, and open reporting. Anyone using the same methods should be able to repeat a study and get a similar result. That property, called replicability, is what gives scientific findings their weight.

The method developed over centuries rather than being invented by one person. In the early 1000s, the scholar Ibn al-Haytham used controlled experiments in his Book of Optics to test theories of vision. Francis Bacon argued for building knowledge from systematic observation in Novum Organum (1620). Galileo Galilei and Isaac Newton joined experiments with mathematics. Modern research has added statistical inference, peer review, and study pre-registration. For a fuller history, see the Stanford Encyclopedia of Philosophy entry on the scientific method.

What Is the First Step of the Scientific Method?

The first step of the scientific method is observation. A researcher notices something that prompts curiosity or doesn't fit what was expected. The observation might be everyday, such as a plant growing faster on one side of a room. It might also be precise, such as a sensor reading that breaks a known pattern.

Some textbooks list "ask a question" as step one and treat observation as a starting condition. Both framings agree on the substance: an investigation starts when someone notices something specific. A vague observation ("plants seem to grow") leads nowhere. A precise one ("this plant grows faster near the south window") gives a researcher something to test.

The Steps of the Scientific Method Explained

Each step below builds on the one before it. Real projects often loop back, but the order describes the logic every study follows.

Step 1. Observation

Observation is noticing something specific that can be measured or described. A clinical researcher might notice that patients on one medication report fewer headaches than expected. A psychologist might notice that students seem to score higher on morning tests. Good observations are concrete enough to point toward a question.

Step 2. Question

The question narrows the observation into something evidence can answer. "Why do students do better in the morning?" is a start. "Does time of day affect scores on a standard working memory task?" is a question a study can test. Before designing anything, researchers review published work to learn what's already known. For frameworks and samples, see how to write a research question and these research question examples.

Step 3. Hypothesis

A hypothesis is a testable, falsifiable prediction that offers an explanation. For example: students score higher on a working memory task in the morning than in the afternoon, because alertness peaks earlier. It names what will be measured and predicts a direction. Most studies also state a null hypothesis, which predicts no difference.

Falsifiability means some possible result could show the hypothesis is wrong. The philosopher Karl Popper made this the dividing line between science and non-science in The Logic of Scientific Discovery (1959). A claim that fits every possible outcome, such as "everything happens for a reason," can't be tested. For 50 samples across fields, see these hypothesis examples.

Step 4. Experiment

The experiment tests the hypothesis by changing one factor and measuring the result. The factor the researcher changes is the independent variable, and the outcome is the dependent variable. Our guide to independent and dependent variables covers both in detail. Other influences are held steady with control variables or evened out by random assignment. For design choices, see experimental research design.

Some fields can't run traditional experiments. Astronomers can't move stars, and epidemiologists can't assign people to smoke. These researchers use observational studies and quasi-experimental designs that compare groups as fairly as the situation allows. The logic stays the same: compare what the hypothesis predicted with what actually happened.

Step 5. Data Collection and Analysis

Analysis turns raw measurements into evidence for or against the hypothesis. Quantitative studies use statistical tests, such as t-tests or ANOVA, to judge whether a difference is larger than chance would produce. A p-value summarizes that judgment, but it doesn't say how large or important an effect is. Qualitative studies analyze interviews or field notes with thematic or interpretive methods. Either way, the analysis must be described clearly enough for someone else to repeat it.

Step 6. Conclusion

The conclusion states whether the data support the hypothesis, and under what limits. If the hypothesis is supported, the researcher reports the effect size and the conditions where it held. If it isn't supported, that's a real result too, because it rules out one explanation. A conclusion never reaches past its data. A study of undergraduates at one university can't show that everyone thinks better in the morning.

Step 7. Communication and Replication

Findings are shared through peer-reviewed journals, conference talks, or open data. Other researchers then try to repeat the study. A single study, however careful, isn't conclusive. Results that replicate across independent teams join the scientific record, and results that don't are revised or dropped. Each finding also raises new questions, which is why the method is often drawn as a cycle.

How Many Steps Are in the Scientific Method? 5, 6, and 7 Compared

The scientific method is taught as 5, 6, or 7 steps, depending on the textbook. The versions differ only in how they group the same activities. Shorter lists merge observation with the question, or analysis with the conclusion. Longer lists give background research or communication a step of their own. The table shows how common versions line up.

If a teacher or textbook asks for "the 5 steps," use the 5-step column. The substance doesn't change, only the grouping.

Scientific Method Examples

The steps make more sense once you see them applied. The first example is a classroom experiment. The second shows the same logic in everyday problem-solving.

Example 1: Sunlight and Plant Growth

1. Observation. A biology teacher notices that plants on her south-facing windowsill grow taller than the same species on the north-facing sill.
2. Question. Does the amount of direct sunlight a plant receives affect how tall it grows over a fixed period?
3. Hypothesis. Plants that get more hours of direct light will grow taller over six weeks, because photosynthesis increases with light.
4. Experiment. She plants 30 seedlings from one seed batch in identical pots and soil. She randomly assigns 10 seedlings each to 4, 8, or 12 hours of light per day. Water and temperature stay the same, and she measures height weekly.
5. Analysis. After six weeks, she compares average final heights across the three groups with a one-way ANOVA.
6. Conclusion. The 12-hour group grew tallest and the 4-hour group grew least, so the hypothesis is supported. The result applies only to this species, this period, and this range of light.
7. Communication. She shares the results with her department. In a research setting, the study would go to a peer-reviewed journal so others could replicate it.

Example 2: A Phone That Won't Charge

The scientific method also works for everyday problems. You observe that your phone won't charge overnight. Your hypothesis is that the cable is faulty. You test it by charging the phone with a different cable and the same outlet. If the phone charges, the evidence supports your hypothesis. If it doesn't, you rule out the cable and test the next explanation, such as the outlet or the charging port.

Inductive and Deductive Reasoning in the Scientific Method

The scientific method uses both kinds of reasoning. Inductive reasoning moves from specific observations to a general idea: several plants grew faster in sunlight, so light may drive growth. Deductive reasoning moves from a general idea to a specific prediction: if light drives growth, plants given 12 hours should outgrow plants given 4.

Induction suggests hypotheses, and deduction turns them into tests. Many studies combine the two, moving between them as results come in. Qualitative research often leans inductive, building theory from data. Quantitative research often leans deductive, testing predictions set in advance. Our guide to quantitative vs qualitative research explains how to choose.

Scientific Method vs. Engineering Design Process

The scientific method and the engineering design process are often taught side by side, but they answer different questions. Science asks how the world works. Engineering asks how to build something that solves a problem. The table compares them.

How the Scientific Method Guards Against Bias

The scientific method is a set of safeguards as much as a sequence of steps. Random assignment spreads existing differences evenly across groups. Blinding keeps participants and researchers from knowing who is in which group, so expectations can't shape results. Pre-registration commits researchers to an analysis plan before they see the data.

Bias still gets through. Selection bias distorts who ends up in a study. Observer bias shapes what researchers record. Publication bias means positive results are more likely to reach print. For the full range, see our research bias guide.

Common Misconceptions About the Scientific Method

• It isn't a rigid recipe. Real research rarely follows the steps in a straight line. Researchers refine hypotheses after pilot studies and revisit observations after seeing data.
• A hypothesis isn't just a guess. It's a specific, testable prediction grounded in prior knowledge. "I think X causes Y" is a guess. "If X causes Y, raising X should raise Y under these conditions" is a hypothesis.
• An unsupported hypothesis isn't a failure. Null results rule out an explanation and save other researchers time. Many journals now encourage publishing them.
• Scientific conclusions are provisional. They're the best-supported answer given current evidence. New evidence can revise them, which is a strength of the method, not a weakness.
• Correlation isn't causation. Two variables can move together because of a shared cause, reverse causation, or chance. Confounding variables are a common culprit. Randomized experiments give the strongest evidence of cause, though well-designed natural experiments can also support causal claims.

How the Scientific Method Shapes a Research Paper

A research paper follows the scientific method section by section. Most empirical papers use the IMRaD structure: Introduction, Methods, Results, and Discussion. The table shows where each step appears.

Theses and dissertations follow the same logic at greater depth, usually across five chapters. Committees and reviewers read them for evidence that the method was applied rigorously and reported transparently. For section-by-section help, see how to write a research paper. Our guides to the methods section and the results section go deeper.

Frequently Asked Questions

Related Research Guides

The research methodology guide for graduate students covers the main design choices researchers face and where the scientific method fits among them.

The experimental research design article walks through true experiments, random assignment, and control groups in more depth than Step 4 above.

The hypothesis examples article gives 50 sample hypotheses across disciplines, with notes on what makes each one testable.

The research bias guide explains the biases the scientific method is designed to guard against and how to reduce each one.

Reviewed by the Editor World editorial team under the direction of Patti Fisher, PhD, founder. Editor World has provided 100% human editing by native English editors since 2010. About Editor World