How to plan a fair test with a working model

A working model becomes more valuable when the learner uses it to answer a question. A fair test does not need complicated equipment. It needs a clear comparison, a method that can be repeated and an honest record of what happened.

1. Ask a question you can test

Start with ‘What happens when…?’ Good questions name both the condition being changed and the result being observed. For example: ‘What happens to turbine speed when the pouring height increases?’ is more useful than ‘How does hydro power work?’ for a short investigation.

2. Name the variables

The independent variable is the one condition you deliberately change. The dependent variable is what you measure or observe. Control variables are the other conditions you try to keep the same so the comparison is meaningful.

With a turbine test, pouring height might be the independent variable, turns in ten seconds the dependent variable, and water volume, turbine position and counting time the controls.

3. Decide what counts as evidence

Choose a measurement before the test begins. It might be time, distance, number of turns, load lifted or the number of successful attempts. If the result is descriptive rather than numerical, create clear categories so different learners record it consistently.

4. Make a prediction

A prediction should include a reason: ‘I think the turbine will turn more times because water released from higher up will reach it with greater speed.’ The prediction may be wrong. Its purpose is to make the learner’s thinking visible before the evidence is known.

5. Repeat the test

One result can be affected by timing, alignment or a small change in technique. Repeat each condition at least three times when possible. Look for a pattern and record unusual runs rather than quietly deleting them.

6. Present the result clearly

A simple table is often enough. Put the independent variable in the first column, each repeated result in the next columns and a summary such as an average in the final column. A graph can help reveal a pattern when both variables are numerical.

7. Separate the claim from the limitations

State what the evidence supports, then identify what the test could not show. A small model may demonstrate a relationship without matching the scale, material or efficiency of a real system. Recognising that limitation is part of good scientific reasoning.

Useful group roles

In a classroom, assign roles that contribute to the investigation: equipment lead, method reader, measurer, recorder and reporter. Rotate roles across tests so participation is not reduced to one learner building while everyone else watches.