Same Science › Planning investigations › Fair tests
What makes a fair test?
A fair test changes only one thing, so any difference in results must come from that change. If one plant gets more light and a bigger pot, and it grows taller, you can't tell which made the difference. Kids learn to spot unfair tests and fix them, and to repeat trials so one odd result doesn't fool them.
- Fair: only the independent variable changes.
- Not fair: two things change, so the result can't be pinned on either.
- Repeat trials to make results more reliable.
Why schools teach it
The heart of planning an experiment
Third through fifth graders plan fair tests in which variables are controlled (3-5-ETS1-3), and middle schoolers design controlled investigations (science practice 3).
Spotting unfair tests trains kids to doubt sloppy claims, a skill that transfers to ads, news and their own science fair projects.
At the kitchen table
What to say, and what to skip
Say this
What's being changed? Is anything else different?
If the results are different, could something else explain it?
How would you make it fair?
Keep the other thing the same.How many times should you test it?
At least three trials.
Skip this
- One trial and done.Repeating trials catches mistakes and odd results.
- Testing everything at once to save time.Then no result means anything.
- Picking the result you expected.Report what the data shows.
See it in action
Fixing: “We tested two paper airplanes. One was printer paper thrown by Sam; one was cardstock thrown by Maya.”
What changed?
the paper AND the throwerWhy that's a problem
a farther flight could be Maya's arm, not the paperFix it
same thrower, same fold, same throw; only the paper changes; three throws each
Ask “what else could explain the difference?”
Free practice sheet
Make a fair test sheet
Loading practice sheet…
Fridge cheat sheets
Science Fridge Cheat Sheets
Every science method in one printable book (36 pages): a cheat sheet page like this guide, plus a practice sheet and its answer key for each, in Classic and handwritten Notebook styles. $8 one time, or all 7 subjects for $29.99. See all prices.
Teacher? This method has a classroom pack on TeachersPayTeachers: an anchor chart, a parent letter and practice sheets with answer keys, $4. See the pack
Parents: tell your child's teacher about this guide
Parents also ask
Why do three trials?
One result can be a fluke. The average of several trials is more trustworthy.
Is a fair test the same as a controlled experiment?
Yes. Elementary classes say fair test; middle school says controlled experiment.
What if you can't keep everything the same?
Keep as much the same as you can and mention the rest as a possible error.
Does a fair test need a control group?
Often. A no-change group (no salt, no fertilizer) gives you something to compare to.