What Is the Fourth Step in the Scientific Method?
Let’s start with something that might surprise you: the fourth step in the scientific method isn’t what most people think it is. You’ve probably memorized the sequence in school—ask a question, do background research, form a hypothesis—but then what? Most textbooks say “test your hypothesis,” and that’s where things get fuzzy. And the fourth step? It’s actually making predictions based on your hypothesis. And no, that’s not the same as testing it.
The Real Order (Spoiler: It’s Not Linear)
Here’s what the scientific method really looks like:
- Ask a question
- Do background research
- Form a hypothesis
- Make predictions
- Test those predictions through experiments
- Analyze the data
- Draw conclusions
- Communicate results
See the difference? In practice, prediction comes before testing. It sounds minor, but it matters. That said, a hypothesis is your educated guess—a starting point. Even so, a prediction is what you expect to see if your hypothesis is right. It’s the bridge between thinking and doing Nothing fancy..
Why Prediction Is the Missing Piece
Most people skip over prediction because it feels obvious. But here’s the thing—you can’t test what you haven’t clearly defined. Predictions force you to be specific. “Of course I’ll test my idea,” they say. They turn vague ideas into measurable outcomes The details matter here..
To give you an idea, if your hypothesis is “plants grow faster with music,” your prediction might be “if plants are exposed to classical music for 2 hours daily, they will show 20% more growth than plants with no sound exposure.” Now you’re not just hoping for the best—you’ve set up a measurable outcome And it works..
Why This Step Actually Matters
Prediction is where science gets real. It’s the moment you commit to something being testable. Without it, you’re just collecting data randomly and hoping it means something. With it? You’re building a roadmap.
It Keeps You Honest
Predictions create accountability. They force you to think through your logic before you start waving the test tube around. If you can’t predict what your results should look like, your hypothesis might be too vague—or worse, unfalsifiable.
Real talk: this is where a lot of amateur science goes off the rails. Someone forms a hypothesis, runs an experiment, and then cherry-picks results that match what they wanted to see. So naturally, predictions prevent that. They give you a benchmark to measure against.
It Makes Experiments Meaningful
Think about it like this: if you’re testing whether fertilizer affects plant growth, you need a control group and an experimental group. But more than that, you need to know what you’re looking for. Your prediction tells you: “If fertilizer works, the treated plants should be taller by week four Practical, not theoretical..
This is the bit that actually matters in practice.
Without that prediction, you’re just watching plants grow and hoping something interesting happens. With it, you’re conducting a focused investigation.
How to Actually Make Good Predictions
This is where most guides fall flat. They’ll tell you to “make predictions,” but not how. So here’s the practical breakdown.
Start With Your Hypothesis
Your prediction should flow directly from your hypothesis. If your hypothesis is “students who study with background music score higher on tests,” your prediction needs to specify:
- What kind of music (classical? Lo-fi?)
- How much music (volume? duration?)
- What kind of test (math? history?)
- By how much you expect scores to improve
See how that narrows things down? That specificity is what makes your experiment valid Turns out it matters..
Make It Measurable
Predictions aren’t about feelings or general trends. They’re about numbers. Here's the thing — “I think it will work better” isn’t a prediction. “I predict a 15% improvement in test scores” is Easy to understand, harder to ignore..
This doesn’t mean everything has to be quantified. Some sciences deal in qualitative data. But even then, you need observable, measurable outcomes. “I predict participants will report feeling more focused” is better than “I think it helps.
Plan for the Unexpected
Here’s a pro tip that most people miss: good predictions include what you’ll do if you’re wrong. That said, if your data doesn’t match your prediction, that’s not failure—it’s information. It means your hypothesis needs refining, or that you’ve discovered something new.
Science isn’t about being right every time. It’s about learning.
Common Mistakes People Make
Let’s be honest—most people screw up this step in one of these ways.
They Skip It Entirely
I’ve seen countless experiments where someone jumps straight to testing without ever stating what they expect to find. That’s backwards. Even so, they collect data, look at it, and then decide what it means. That’s not science—that’s storytelling with numbers.
They Make Predictions Too Broad
“I think the medicine will help people feel better” isn’t a prediction. Because of that, “I predict patients will report a 30% reduction in symptoms after two weeks” is. The first one could mean anything. The second gives you something concrete to measure.
They Ignore Variables
Every experiment has variables. Some you control, some you don’t. Good predictions account for this. They acknowledge what you’re changing and what you’re keeping constant.
For instance: “If temperature affects reaction rate, then increasing the temperature from 20°C to 40°C will double the reaction speed, assuming all other conditions remain the same.”
Notice how that sets up the experiment perfectly?
What Actually Works in Practice
After years of reading and testing different approaches, here’s what I’ve found works:
Write Down Your Prediction Before You Start
Seriously. Even if it’s just a sentence. Write it down. This simple act forces you to think through your logic and makes you more likely to stick to your plan when you’re in the middle of messy data collection Small thing, real impact..
Use If-Then Language
Predictions work best when framed as “If X, then Y.” This structure keeps your thinking clear and your experiments focused. It also makes it easier to spot when your results don’t match your expectations Simple, but easy to overlook. Simple as that..
Keep It Realistic
Don’t make predictions so ambitious that they’re impossible to test. Plus, if your hypothesis is “this new drug cures cancer,” your prediction better include a realistic timeline and measurable endpoints. Otherwise, you’re setting yourself up for disappointment—or worse, manipulation.
Test Your Prediction, Not Just Your Hypothesis
This is subtle but important. If the predictions hold up, your hypothesis gains support. Now, you test your hypothesis by testing your predictions. If they don’t, your hypothesis needs work Easy to understand, harder to ignore..
Frequently Asked Questions
Is making predictions part of the hypothesis?
Sort of. Your hypothesis is the broader statement (“If I do X, then Y will happen”). That's why your prediction is the specific, measurable version of that (“I expect Y to increase by 20%”). The prediction is a component of testing your hypothesis.
Can you do science without making predictions?
In a pinch, maybe. Predictions are what make experiments controlled and repeatable. But it won’t be rigorous science. Without them, you’re just observing—and that’s not the same as investigating.
What if my prediction is wrong?
Then you learn something. Wrong predictions are valuable—they tell you your hypothesis needs adjusting. That’s how science progresses, one wrong prediction at a time.
Do all sciences use the same prediction format?
The principle is the same across fields, but the specifics vary. On the flip side, physics might predict numerical values. Consider this: biology might predict observable traits. Social sciences might predict behavioral trends. The key is always being specific and measurable.
How specific should a prediction be?
Specific enough that you could, in theory, prove it wrong. Vague predictions can’t be tested, and untestable ideas aren’t scientific.
Wrapping It Up
So there you have it—the fourth step that most people forget: making predictions. It’s not flashy. It doesn’t require expensive equipment or fancy theories. But it’s essential.
Prediction is what turns a good idea into a real experiment. In practice, it’s the moment you shift from wondering to testing. From hoping to knowing.
Most importantly, it keeps you honest. When you’ve clearly stated what you expect to find, you can’t later decide that any result counts as a win. Science demands better than that—and so do you, if you want to do it right Not complicated — just consistent. Which is the point..
Most guides skip this. Don't.
The next time you design an experiment, don’t skip this step. On the flip side, write down what you expect to see. Make it measurable. Make it specific. And then go test it Simple as that..
That’s how
That’s how you separate science from storytelling. That’s how you turn curiosity into knowledge. And that’s how you see to it that when the data comes in—whatever it shows—you’ll have the clarity to understand what it actually means.
The scientific method isn’t a rigid checklist to be ticked off mindlessly. So observation grounds you. It’s a discipline of thought. Questions focus you. But predictions? Day to day, hypotheses guide you. Each step constrains the next, building a structure that can withstand scrutiny. Predictions commit you.
And that commitment is the whole point.
So write the prediction. Put it in your lab notebook, your pre-registration, your grant proposal. Then run the experiment. Make it public if you can. Let the world tell you whether you were right Small thing, real impact. That alone is useful..
Because in the end, science isn’t about being right. Because of that, it’s about finding out. And a clear, testable prediction is the only tool that lets you do that with integrity The details matter here..
