Making science means following a structured process to answer a question about the natural world. It starts with curiosity about something you’ve observed, moves through careful testing, and ends with a conclusion backed by evidence. Whether you’re designing a school project or trying to understand how professional research works, the core process is the same.
Start With a Question You Can Test
Every piece of science begins with noticing something and wondering why it happens. Maybe you’ve noticed that plants on one side of your yard grow taller, or that you sleep worse on nights you use your phone late. The first step is turning that curiosity into a specific, testable question: “Do plants grow taller with more sunlight?” or “Does screen time before bed reduce sleep quality?”
That testable question is your hypothesis. A good hypothesis makes a prediction that could be proven wrong. This is the single most important quality that separates science from guesswork. If no possible result could disprove your idea, it isn’t scientific. The philosopher Karl Popper identified this principle of falsifiability as the clearest way to distinguish science from pseudoscience. For example, “this crystal has good energy” isn’t scientific because there’s no observation that could show it’s false. But “watering a plant twice a day will make it grow faster than watering it once” is scientific because you can measure the result and be wrong.
Learn What Others Have Already Found
Before you design an experiment, find out what’s already known. Someone may have tested the same question, and their work can help you refine your approach or avoid mistakes. Review existing research, textbooks, or reliable sources on your topic. Even if someone has already published results, you can still repeat their experiment to see if you get the same answer. In fact, replication is one of the most valuable things in science. A major meta-analysis of biomedical research found that only about 50% of preclinical studies could be successfully reproduced, which means there’s real value in testing things again.
Design an Experiment That Controls for Bias
The core of making science is the experiment itself, and a well-designed one has three key ingredients: variables you change, variables you measure, and controls you keep the same. If you’re testing whether fertilizer helps plants grow, the fertilizer is what you change (the independent variable), plant height is what you measure (the dependent variable), and everything else, like sunlight, water, and soil type, stays constant (the controls).
The gold standard in professional research is the randomized controlled trial. Participants are randomly assigned to either the group receiving the treatment or a comparison group, which removes the influence of personal characteristics that might skew results. In many studies, neither the participants nor the researchers know who’s in which group, a technique called blinding, which prevents expectations from shaping the outcome. You don’t need this level of complexity for a school project, but the principle applies everywhere: change only one thing at a time, and keep everything else the same so you know what actually caused your result.
Write out your procedure step by step before you begin. This forces you to think through practical details and makes it possible for someone else to repeat exactly what you did.
Collect and Analyze Your Data Carefully
Run your experiment and record everything. Write down not just the results you expected to track, but anything unusual that happens along the way. Precise, honest record-keeping is the backbone of credible science.
Once you have your data, look for patterns. This might mean calculating averages, creating graphs, or running statistical tests depending on your level. In professional research, scientists typically use a threshold called a p-value to decide whether their results are meaningful or just due to chance. A p-value below 0.05 (meaning there’s less than a 5% probability the result happened randomly) is the most common standard for calling a finding “statistically significant.” For simpler projects, a clear visual pattern in a graph or a consistent difference between your groups can tell the story.
Draw Honest Conclusions
Compare your results to your original hypothesis. Did the data support your prediction, or not? Both outcomes are valid science. A hypothesis that turns out to be wrong still produces useful knowledge because it rules out one explanation and points you toward better questions.
Be honest about limitations. Were there factors you couldn’t control? Was your sample size small? Could measurement errors have affected the outcome? Professional scientists address these questions in every published paper, and acknowledging weaknesses actually makes your work stronger, not weaker. If your results didn’t support your hypothesis, suggest alternative explanations and what you’d test next.
How Hypothesis, Theory, and Law Differ
Understanding these terms helps you see where your work fits into the bigger picture. A hypothesis is a tentative, testable explanation. It can be easily changed based on new evidence. A scientific theory is something much more established: a comprehensive explanation of a natural phenomenon supported by a large body of evidence gathered over time. Theories like evolution or germ theory aren’t guesses. They are reliable accounts of how the world works, confirmed through repeated observation and experimentation.
A scientific law describes a consistent pattern in nature, often expressed as an equation (like the law of gravity), but doesn’t explain why the pattern exists. A theory explains the why. Both are well-supported by evidence, but they serve different purposes.
What Happens in Professional Science
If you’re curious how the process scales up beyond a classroom, professional science adds several layers. Researchers typically apply for funding through agencies like the National Institutes of Health, the largest public funder of biomedical research in the world. Grant applications go through two levels of review: first by other scientists who evaluate the proposal’s quality, then by an advisory council.
Any research involving human participants must follow strict ethical principles. The three core requirements, established in a foundational document called the Belmont Report, are respect for persons (people must give informed, voluntary consent), beneficence (the research must minimize harm and maximize benefit), and justice (the burdens and benefits of research must be distributed fairly). An independent review board evaluates whether the risks of any study are justified before it can begin.
Once a study is complete, researchers write up their findings and submit them to a scientific journal. The manuscript goes through peer review, where other experts in the field evaluate the methods, analysis, and conclusions. A single review takes about 6 hours on average, but the full process from submission to publication typically takes 12 to 14 weeks in medical and natural science journals, and can stretch to 25 weeks or more in fields like economics. About one-third of papers sit for two weeks before even receiving an initial editorial decision.
Published data increasingly follows what are known as FAIR principles: it should be findable, accessible, interoperable (usable across different systems), and reusable by other researchers. This transparency is what allows science to self-correct over time, as others can check the original data and test the same questions independently.
Making Science at Any Level
You don’t need a lab or a degree to make science. You need a question, a fair test, honest data, and a willingness to be wrong. A kid measuring how far rubber bands of different thicknesses stretch is doing the same fundamental thing as a researcher running a clinical trial: isolating a variable, measuring an outcome, and drawing a conclusion from evidence. The scale changes, the rigor increases, but the logic is identical.
Start small. Pick something you’re genuinely curious about, write down a prediction, test it as carefully as you can, and see what the data tells you. That’s how you make science.