How to use logic and evidence to judge a scientific claim
TEAS 7 outline: Apply logic and evidence to a scientific explanation
- A hypothesis guesses, a theory explains, and a law describes.
- A good hypothesis must be testable and falsifiable.
- Evidence can support or refute a hypothesis, but never proves it.
- Correlation alone does not show cause; a third variable may drive both.
- Inductive goes from cases to a rule; deductive goes from rule to prediction.
Key facts
- A hypothesis is a tentative, testable explanation; it must be falsifiable, meaning some possible result could show it is wrong.
- A theory is a well-substantiated, testable explanation of a broad set of observations; a law summarizes many observations and describes or predicts what happens.
- Results can support or refute a hypothesis, but science does not claim to prove a hypothesis; new evidence can always change our understanding.
Show all 6 key factsShow fewer
- A correlation means two variables change together; it does not by itself show that one causes the other, because a third (confounding) variable may drive both.
- A controlled experiment, ideally with random assignment to groups, is the way to test whether one variable causes a change in another.
- Inductive reasoning moves from specific observations to a general conclusion, which may turn out wrong; deductive reasoning moves from a general principle to a specific prediction.
Quick check1 question · 2 more at the end
Which statement correctly compares a scientific law with a scientific theory?
Correct answer: B.
A scientific law summarizes a large number of observations and describes or predicts what will happen. A theory is a well-substantiated, testable explanation of why a broad set of observations occur. They are different kinds of statements, so a theory does not turn into a law, and neither one is ever considered permanently proven; both can be revised if new evidence requires it.
A This is the common myth that theories graduate into laws; a theory explains and a law describes, so one does not become the other.
B Correct: laws describe or predict natural behavior, and theories offer well-tested explanations for it.
C An untested idea is closer to a hypothesis, and science does not label any hypothesis or theory as proven.
D This reverses the roles and shrinks a theory to one observation; theories explain broad sets of observations.
Full lesson
Each part starts with its main idea. Open a part when you want the details.
The short answer: A hypothesis is a testable, tentative idea; a theory is a well-supported explanation of many observations; a law describes or predicts what happens. Evidence can support or refute a hypothesis but never proves it. Correlation alone does not show cause. Inductive reasoning goes from specific cases to a general rule; deductive goes from a rule to a prediction.
1 What is the difference between a hypothesis, a theory and a law?The key is to ask what each one does: guess, explain, or describe. 1 min
The key is to ask what each one does: guess, explain, or describe.
| Term | What it is | What it does | Example |
|---|---|---|---|
| Hypothesis | A tentative explanation that can be tested | Proposes a possible answer to one question | "My phone won't charge because the cable is damaged." |
| Theory | A well-substantiated explanation backed by many tests | Explains why a wide range of observations happen | Cell theory: all living things are made of cells, and new cells come from existing cells |
| Law | A summary of a huge number of observations | Describes or predicts what happens | Boyle's law: at constant temperature, a gas's volume goes down as pressure goes up |
A theory does not "turn into" a law once it is tested enough. A law tells you what will happen; a theory tells you why. Both can be revised if new data demand it.
A good hypothesis must be testable and falsifiable: there has to be some possible result that would show it is wrong. "Invisible spirits drain my battery, but they can never be detected" is not a scientific hypothesis, because no test could ever rule it out.
2 Can an experiment prove a hypothesis?Results can support a hypothesis or refute (reject) it, but science does not claim to prove anything. 1 min
No. Results can support a hypothesis or refute (reject) it, but science does not claim to prove anything.
Try a mini-experiment. Your phone won't charge, and your hypothesis is "the cable is damaged." Your prediction: if the cable is damaged, then a different phone plugged into the same cable won't charge either.
- The other phone charges fine. That result refutes your hypothesis, so you propose a new one (maybe the phone's charging port is dirty).
- The other phone also fails. That result supports your hypothesis, but it doesn't prove it. The wall outlet could be dead. A second test, using the cable in a different outlet, would rule that out.
Notice the logic: one solid result that contradicts a prediction is enough to reject a hypothesis, while a matching result only makes it more believable.
3 What makes evidence count?Good evidence is a result that would come out differently depending on whether the claim is true. 1 min
Good evidence is a result that would come out differently depending on whether the claim is true. When two explanations compete, look for the test where they predict different outcomes.
Say Lena thinks a houseplant is drooping from too little water, and Omar thinks it's from too little light. Watering the plant and moving it to a sunny window at the same time is a weak test: if it perks up, you can't tell which change helped. Watering it while leaving it in the same spot is a strong test, because the two explanations predict different results.
Also ask whether the evidence is empirical: something observed and measured that others could check again, not a feeling or a single story.
4 Does a correlation show cause and effect?A correlation means two variables change together. 1 min
A correlation means two variables change together. It does not, by itself, show that one causes the other.
Imagine a city finds that on days when more sunscreen is sold, more people go to the beach and more lifeguard rescues happen. Does sunscreen cause rescues? No. A third factor, hot sunny weather, drives both. This hidden factor is called a confounding (or lurking) variable.
To test cause and effect, you need a controlled experiment. Change one variable on purpose, keep everything else the same, and ideally use random assignment so the groups start out alike. Then a difference between the groups can be traced to the one thing you changed.
5 What's the difference between inductive and deductive reasoning?Induction builds a general idea from data, and deduction turns that idea into a testable prediction. 1 min
| Inductive | Deductive | |
|---|---|---|
| Direction | Specific observations → general conclusion | General principle → specific prediction |
| Example | "Every one of the 40 beetles I caught in this field was black, so beetles here are probably black." | "A gas expands when heated at constant pressure. Helium is a gas. So a helium balloon should expand when warmed." |
| Certainty | Can be wrong; the next beetle could be green | If the principle is true and applies, the prediction follows |
Scientists use both. Induction builds a general idea from data, and deduction turns that idea into a testable prediction.
6 How do you evaluate a scientific claim? 1 min
Run any claim through a short checklist:
- Is it testable? Could any result prove it wrong?
- How many cases? A handful of people can't settle it.
- Was there a comparison group? Without one, you can't tell what would have happened anyway.
- Did people know what they got? Expectations alone can change how people feel, which is why studies use placebos and blinding.
- Correlation or experiment? Only an experiment can show cause.
- Has it been checked by others? Peer review and repeated studies make a finding more trustworthy.
Common mistakes
The traps students fall into most. Read these before you practice.
Watch out: A theory does not graduate into a law after enough testing; a theory explains and a law describes, so they are different kinds of statements.
Watch out: Data that match a prediction support a hypothesis, they do not prove it.
Watch out: Two things rising together is a correlation, not proof that one causes the other.
Watch out: Inductive goes specific to general and deductive goes general to specific, not the other way around.
Watch out: Results that would happen whether the claim is true or false are not useful evidence for either side.
How to remember it
Memory hook: "Hypothesis guesses, theory explains, law describes." And "IN-duction goes IN to the big picture": from small details up to a general rule.
Self-check:
- Explain why "the theory became a law" is a mistake.
- Name the confounding variable if ice skating rentals and hot cocoa sales rise together.
- Turn "plants need light to stay green" into an if-then prediction, then describe one result that would refute it.