Guide for parents · September 2026
Fair Tests With Guided Practice
What a fair test is in Year 7 science: independent, dependent and control variables, repeats and anomalies, and how to check a plan at home.
A fair test is a definition, not an attitude
The phrase sounds like it is about playing by the rules, and children often hear it that way. It is not. A fair test is an investigation designed so that only one thing is allowed to differ between the trials, which means any difference in the results can be blamed on that one thing and nothing else. Change two things and the experiment can no longer answer its own question, however carefully the readings were taken.
That is why teachers keep returning to it from Year 7 onwards. It is the piece of reasoning that makes every later practical worth doing, and it is the piece children get the least practice at once the lesson ends.
The three kinds of variable
Everything in a fair test falls into one of three roles, and a child who can sort them can plan almost any investigation.
| Role | Plain meaning | In an insulation investigation |
|---|---|---|
| Independent variable | The one thing you deliberately change | The material wrapped round the beaker |
| Dependent variable | The thing you measure to see what happened | The water temperature after ten minutes |
| Control variables | Everything you deliberately keep the same | Starting temperature, volume of water, beaker, thickness of wrapping, room |
Two habits make the difference. The independent variable should have planned values written down before anything is touched, not chosen as you go. The control variables should be listed, because a list is testable and a vague intention is not.
There is also a trap in the vocabulary. A control variable is something held constant, while a control experiment is a separate run with no treatment at all — the beaker with nothing wrapped round it. They share a word and mean different things, and mixing them up loses marks in exactly the questions that were meant to be easy.
The three questions that check any plan
You do not need any science to test whether a plan is sound. Ask your child what they are changing, what they are measuring, and what they are keeping the same. If all three come back clearly, the design works. If the first two blur together, or the third is "everything", that is the conversation to have, and it is more useful than checking the results table.
The commonest failure is an independent variable that quietly carries a second change with it. Comparing a thick wool wrap against a thin foil wrap changes the material and the thickness at once, so whatever the thermometer says, the investigation cannot separate the two.
Repeats, means and anomalies
One reading is a number. Three readings are evidence. Repeating a measurement and taking the mean reduces the effect of small errors in timing, reading a scale or starting conditions, and it also makes anomalies visible — a value that sits well away from the others in its own group.
An anomaly is not simply a result you dislike. It is identified against the other repeats at the same setting of the independent variable, and it is usually left in the table, marked, and excluded from the mean, with a sentence saying so. Children who silently delete it lose the mark that was there for noticing it.
Range and interval belong here too. Five settings spread across a sensible range beat two settings at the extremes, because two points cannot show whether a relationship is a straight line or a curve.
Where the marks actually go
Marks in this part of science are rarely lost on science. They go missing on the surrounding skills.
| What the question wants | What it is really testing |
|---|---|
| "Name the independent variable" | Whether the roles were sorted before the practical started |
| "Suggest two control variables" | Whether the child can name specifics rather than say "keep it fair" |
| "Describe the pattern in the results" | Reading a trend from a table or graph, not quoting single values |
| "Explain why the results are unreliable" | Repeats, anomalies and measurement, not the topic |
| Calculating a mean, converting units | Arithmetic and place value |
| Choosing a bar chart or a line graph | Whether the independent variable is categoric or continuous |
That last row catches a lot of children. Different materials are categories, so they get bars. Different masses or temperatures are continuous, so they get a line graph with a drawn line of best fit.
Describe, explain and conclude are three different jobs
A conclusion has to stay inside the data. If insulation was tested for ten minutes on three materials, the conclusion can say which of those three kept the water warmest over ten minutes. It cannot say which is best in a house, or what would happen after an hour, or that the result would hold for any thickness.
Describing means saying what the pattern is: as the thickness increased, the temperature drop got smaller. Explaining means saying why, using science. Children who answer the describe question with a reason, or the explain question with a restatement, lose marks while knowing the material perfectly well.
What Pobbi can honestly help with here
Not much of the practical work, and it is worth being direct about that. There is no equipment, no lab, no exam board specification and no substitute for having actually run an investigation and watched it go wrong. Working scientifically is learned by doing it in school.
What does transfer is the numerical layer underneath: means, scales, units, converting between them, reading a graph and turning a table into a trend. The maths check tests those underlying number skills directly, and if a science question is failing on arithmetic rather than on science, that is worth knowing before anyone books science help.
Questions parents ask about fair tests
My child understands the experiment but writes poor conclusions.
Almost always overreach. Ask them to underline every claim in their conclusion and point to the row in their results table that supports it. Anything left unmarked comes out.
The school keeps saying "improve the validity". What does that mean?
Validity is about whether the investigation measures what it claims to measure — usually a control variable that was not controlled. Reliability is about whether the same readings would come back if you did it again, which is what repeats address. They are separate criticisms and the mark scheme treats them separately.
Is it worth doing experiments at home?
The planning is worth more than the equipment. Pick anything domestic, such as which washing-up liquid makes the longest-lasting bubbles, and spend the time on the three questions rather than on the result.
Does this only matter in Year 7?
No. The same reasoning carries through Key Stage 3 and into GCSE required practicals, where the variables and the reliability questions are worth a substantial share of the marks.