Science // Scientific Inquiry

Fair testing (age 9+)

Use test results to make predictions and set up further comparative and fair tests to investigate new questions

Ages
9 to 11
Type
Thinking skill
Builds on
3 topics
Leads to
1 topic
Standards
1 code

Quick check

After getting results from one experiment, can the learner use the pattern to predict what would happen in a new situation and then design a follow-up test to check?

Ask this to see whether the topic is secure.

Topic // Mastery

What mastery looks like

What a learner can do once this topic is secure.
  • Use results from an investigation to make a specific, testable prediction
  • Design a follow-up test to verify the prediction
  • Explain the reasoning linking the original results to the new prediction

Topic // Map

Where this fits

What to learn first and what this topic opens up, with the reason for each link. Required links come first.
  • RequiredMust come first
  • HelpfulHelps, but not essential

Before this

3 topics
  • Drawing conclusions from evidenceRequired

    Scientific Inquiry // Ages 7 to 9

    Must draw conclusions and make predictions before using results to set up further tests

  • Controlling variablesRequired

    Scientific Inquiry // Ages 9 to 11

    Must plan enquiries independently before designing follow-up investigations from predictions

  • Describing Rules & PatternsHelpful

    Learning to Learn // Learning to Learn // Ages 8 to 9

    Using test results to make predictions and set up further tests is the science form of the universal generalisation habit — describing a pattern as a rule and testing it further

You are here

Fair testing (age 9+)

Ages 9 to 11 // Thinking skill

Leads to

1 topic

Topic // Research-informed

Teaching notes

From the Marble Skill Taxonomy. These notes have not been validated in classrooms, so treat them as a starting point for your own judgement.

The approach

For ages 9–11, teach the extrapolation-and-test cycle in fair testing: read the trend in an existing set of results, use it to make a specific, testable prediction for a NEW untested value or condition, design a fresh fair test to check that prediction, and explicitly state the reasoning that connects the original data to the prediction. The distinctive skill is turning a one-off experiment into an iterative line of enquiry, while learning that an extrapolated prediction is a claim to be tested, not a proven fact.

Common misconceptions

  • Making a prediction as a random guess rather than reading it off the existing trend in the data.
  • Treating an extrapolated prediction as a proven fact instead of a claim that still needs its own fair test.
  • Assuming a pattern continues indefinitely beyond the tested range (e.g. that a trend never levels off or reverses) without checking.
  • Skipping the reasoning link — stating a prediction without connecting it to the specific results that motivate it.
  • Believing an experiment 'ends' at its conclusion, rather than that good results generate the next testable question.

Useful representations

  • A results table/graph with the trend arrow extended to a predicted new value (interpolation vs extrapolation marked)
  • The sentence frame 'I predict ___ because the results showed ___'
  • An enquiry cycle: question → fair test → results → prediction → new fair test
  • A line graph where pupils extend the line of best fit to read off a prediction
  • A prediction-then-test comparison card (predicted value vs measured value)

Teaching practices

  • Teach prediction as reading the trend, not guessing: pupils point to the pattern in the existing results ('as X increased, Y increased') and extend it to a new, unmeasured value — the prediction is evidence-based, not a hunch.
  • Distinguish interpolation (a value between tested points) from extrapolation (beyond them) and flag that extrapolated predictions are less certain and precisely why they must be TESTED, not assumed.
  • Make the reasoning link explicit and spoken/written: 'I predict… BECAUSE the results showed…' — the justification is the assessed skill, not just the prediction itself.
  • Design the follow-up as a genuine fair test in its own right (one variable changed, the rest controlled, repeat readings) so the new question is answered validly.
  • Model science as an iterative cycle — results raise the next question — rather than a single linear experiment that 'finishes' at a conclusion.

Activities to try

  1. 1

    Predict a new point, then run the test to check it

    Why it works: Good data lets you make an evidence-based prediction for conditions you have not yet tested, but an extrapolated prediction is a claim that must be checked with its own fair test — the pattern may not continue indefinitely, so science is a loop where each result raises the next question.

    1. Study a graph of results from a completed fair test (e.g. sugar dissolving faster in warmer water, or distance a car rolls from three ramp heights).
    2. Point to the trend in the data in words ('as temperature increased, dissolving got faster') before predicting anything.
    3. Extend the trend line to a temperature/height you never tested and lock in a specific predicted value for that new condition.
    4. Type or say the reasoning: 'I predict ___ BECAUSE the results showed ___' — before any test is run.
    5. Run a fresh fair test at that new value, keeping everything else the same and taking repeat readings, and place the measured point on the graph next to your prediction.
  2. 2

    Interpolation vs extrapolation: where is the prediction, and how sure are we?

    Why it works: Predictions between tested points are generally more secure than predictions beyond the tested range; naming which kind a prediction is tells you how much it needs checking, and an extrapolated prediction especially must be tested because the pattern might not continue.

    1. On a results graph, mark predictions that fall BETWEEN tested points (interpolation) and predictions that fall BEYOND the tested range (extrapolation).
    2. For each prediction, state whether it is interpolation or extrapolation and rate how certain you can be, explaining why extrapolated predictions carry more uncertainty.
    3. Choose which of several written predictions is best justified by the given data, and explain what makes the others weaker (guess, restated data, or unsupported leap).
    4. Sequence the enquiry cards into the cycle: question → fair test → results → prediction → new fair test.
    5. Say why an extrapolated prediction in particular must be tested rather than assumed.
  3. 3

    Justify and defend your prediction in dialogue

    Why it works: Articulating and defending the evidence-to-prediction reasoning out loud is the core assessed skill; a prediction without its justification and a fair plan to check it is incomplete.

    1. Answer a mentor's question: 'What does this data make you predict for a value you haven't tried — and why that value?'
    2. Explain how your prediction follows FROM the observed pattern, not from a hunch.
    3. Describe how you would set up the follow-up test so it is fair and actually checks your prediction (one variable changed, the rest controlled, repeat readings).
    4. Respond to the challenge: 'What if the trend stops or levels off past a certain point — how would your test reveal that?'
    5. Explain how making an evidence-based prediction differs from simply repeating what you already found.

Topic // Standards

Curriculum standards

Codes for the curriculum statements this topic maps to. Each publisher owns its standard text.
The national curriculum in England: Key stages 1 and 2 framework document
KS2U.Sci.WS.4

Science // Scientific Inquiry

More in Scientific Inquiry

All 30 topics