The Scientific Method
The scientific method in four steps
| Step | What happens |
|---|---|
| Hypothesis | State a falsifiable claim about the situation |
| Design | Plan how to test the claim |
| Experiment | Run the test |
| Evaluation | Check whether the data supports or falsifies the hypothesis |
Why the method matters
The scientific method is possibly the single most important cognitive innovation in history. Its purpose is to falsify theoretical hypotheses through experience.
Statistical analysis
When experimentation is not possible (as with human beings in many cases), statistical analysis is the alternative. It establishes relations among variables based on observation of samples, without changing what is being observed.
The teacher’s translation
A teacher does not need a research lab. A small classroom test, run with a clear hypothesis and an honest evaluation, applies the same logic.
A teacher hears that group work always produces better learning. The teacher tries it for a week and the results are mixed. Did the experiment fail, or was the original claim wrong? The scientific method is the toolkit for answering this kind of question rigorously.
The point of the method is not to make a teacher into a scientist. The point is to give the teacher a way of testing claims that produces real information rather than personal impression.
What the scientific method is
The purpose of the scientific method is to falsify theoretical hypotheses through experience. It is possibly the single most important cognitive innovation in history.
The phrase “falsify through experience” is key. The method does not try to prove things true. It tries to prove them false. A claim that survives many honest attempts to prove it false earns provisional acceptance. A claim that is falsified is set aside.
This shift, from proving true to attempting to falsify, is the core insight. It changes how a teacher thinks about a method or a claim.
The four steps
The scientific method has four steps.
Hypothesis
State a claim about the situation in falsifiable form. The claim should be clear enough that some real-world evidence could prove it wrong.
A non-falsifiable claim like “good teaching matters” cannot be tested. There is no evidence that would prove it wrong. A falsifiable claim like “students who receive feedback within twenty-four hours score higher on the next test than students who receive feedback after a week” can be tested.
Design
Plan the test. The design specifies what variable will be changed, what will be kept constant, what data will be collected, and what counts as falsifying the hypothesis.
A poor design produces results that cannot tell whether the hypothesis is supported. A good design produces a clear answer one way or the other.
Experiment
Run the test as designed. Collect the data. The discipline here is the same as in classroom experimentation: do not change the design partway through.
Evaluation
Compare the data to the hypothesis. Did the data support it, or did the data falsify it? Either answer is useful.
A hypothesis that survives the test is not “proved”. It is provisionally accepted. A different test, with different conditions, might still falsify it. The teacher who understands this stays open to revising even successful hypotheses when new evidence appears.
A hypothesis that is falsified is more informative than people often realise. It removes one wrong idea, which is one fewer wrong idea cluttering the teacher’s practice.
Why falsifiability matters
A claim that cannot be falsified is not a hypothesis. It is a statement of belief. Beliefs may be true, but they cannot be tested.
Many teaching slogans are non-falsifiable. “We value the whole child.” “Every child can learn.” “Reflective practice produces better teachers.” Each of these is comforting and probably has truth in it, but none of them can be falsified, because there is no evidence that would prove them wrong. They survive every test because they were not designed to be tested.
A reflective teacher learns to spot non-falsifiable claims and to translate them into testable ones when possible.
| Slogan | Possible testable version |
|---|---|
| Every child can learn | Within this class, every student can master the unit’s three core skills with appropriate scaffolding within four weeks |
| Reflective practice helps teachers | Teachers who write a structured reflection weekly produce more varied lesson plans over a term than teachers who do not |
| We value the whole child | Our students report higher than school-mean satisfaction in the bi-annual survey |
The translated versions are not perfect. They are testable. The original versions are not.
Statistical analysis
Experimentation works when you can change a variable and watch the effect. Many situations in teaching do not allow this. You cannot run a real experiment on whether grade 8 students learn better in winter than in summer; the seasons come once each. You cannot run an experiment on whether students from particular backgrounds engage with a topic; doing so might be unethical.
Statistical analysis is the alternative.
Statistical analysis establishes relations among variables, or verifies a model, based on the observation of samples of a universe. The data come from observation of reality without changing it.
Statistical analysis is the key when experimentation is not possible, particularly when the object of research is human beings.
For a teacher, statistical analysis is mostly a method for working with existing data. Test scores. Attendance records. Survey results. The teacher does not change anything; they look at what is already there and look for patterns.
A simple example
A teacher wonders whether students who attend the morning section perform differently from those who attend the afternoon section. Running an experiment is not possible (sections are assigned, not chosen). Statistical analysis can compare the test scores of the two sections over a term and see whether there is a real difference.
If there is, the next question is why. The “why” cannot be answered by statistics alone. It requires further investigation.
The teacher’s version
A teacher does not need a laboratory or a statistician. A small, careful application of the scientific method works in a classroom.
A practical version of the method:
- Hypothesis. “If I open the lesson with a question, more students will participate in the first ten minutes than when I open with a statement.”
- Design. Run the question opening for a week and the statement opening for the next week. Count participating students in the first ten minutes.
- Experiment. Do as planned. Keep notes.
- Evaluation. Compare the counts. If question-opening produced more participation, the hypothesis is supported. If not, it is falsified or unclear.
This is not publishable research. It is sound teacher inquiry. The discipline of running it produces results worth more than weeks of unstructured impression.
To falsify hypotheses through experience
The method does not try to prove claims true. It tries to prove them false. A claim that survives honest attempts at falsification earns provisional acceptance. A claim that is falsified is more informative than it might seem: it removes one wrong idea from practice.
When the scientific method is the wrong tool
Three signs that the scientific method, even in its teacher version, is not the right tool:
- The variable is not measurable. Some teaching effects are real but resist measurement at small scale. A method that requires measurement cannot help.
- The time horizon is too long. Some effects only appear over years. A teacher’s classroom inquiry usually cannot run that long.
- The cost of falsifying is too high. If the experiment requires giving one section a method likely to be worse, the ethical cost may outweigh the information gained.
For these cases, other methods are needed. Statistical analysis on existing data, the SECI cycle for tacit knowledge, design and modeling all sit alongside experimentation as parts of a wider toolkit.