Science fair guide
How to write a hypothesis — and pick your variables
A science fair project stands or falls on one sentence and three lists. The sentence is your hypothesis. The lists are your variables. Get them right before you touch any equipment and the rest of the project — procedure, data table, conclusion — almost writes itself.
The hypothesis: if… then… because…
A hypothesis is a prediction you can test, written before you run the experiment. Here is the difference that makes:
Weak
“Plants like music.”
Testable
“If seedlings hear music for six hours a day, then they will grow taller in two weeks than seedlings in silence, because sound vibration may stimulate growth processes.”
The strong version has a specific change, a measurable outcome, a timeframe, and a reason. Being wrong is fine — judges would rather see a clear hypothesis the data rejects than a vague one that can never be wrong.
One change only
If your “if” contains two changes, you no longer know which one caused the result.
A number must come out of it
Taller, faster, heavier, warmer, more — something you can put in a table. “Better” is not a measurement.
Write it before you test
Changing the hypothesis afterwards to match your results is the one thing judges always spot.
The variables: one changes, one is measured, the rest stand still
Independent variable
The one thing you change on purpose. There is only one.
Dependent variable
The thing you measure to see the effect. It depends on the independent variable.
Controlled variables
Everything else you deliberately keep the same, so the only possible cause of a difference is your one change.
Worked example: does sugar change how much gas yeast produces?
A real project from the library — Does the amount of sugar affect how much carbon dioxide yeast produces?
- Hypothesis: If I increase the sugar, then the yeast will produce more carbon dioxide in 30 minutes, because sugar is the yeast's food.
- Independent variable: amount of sugar (0 g, 5 g, 10 g, 15 g).
- Dependent variable: carbon dioxide produced, measured by balloon circumference after 30 minutes.
- Controlled variables: same yeast amount, same water temperature, same bottle size, same 30 minutes, same room.
That project's page lists its variables exactly this way — every project in the library does, because the workspace won't let you skip them.
Worked example: which bridge design holds the most weight?
- Independent variable: the bridge design — same span, same materials budget. (See Which bridge design supports the most weight?)
- Dependent variable: maximum load before collapse, in grams.
- Controlled variables: same span length, same material type, same loading method — weight added at the centre, in the same increments.
Common mistakes judges see every year
The demo disguised as an experiment
A volcano that erupts is a demonstration. It becomes an experiment when one thing changes — say, the amount of baking soda — and something gets measured, like eruption height or duration.
Two independent variables
Testing a new soil and a new light level at once tells you nothing about either.
One trial
A single result can be luck. Three trials at each setting is the minimum judges look for; the library's data tables are built for repeated trials.
Measuring by eye
“It looked faster” loses to “it took 12 seconds instead of 19” every single time.
Pick any project in the library and its page already states the question, the variables and a data table with the columns you need. Start from one of those, adjust the independent variable to make it your own question, and your hypothesis is halfway written.
Questions, answered
What is a hypothesis in a science fair project?
A testable prediction written before you experiment, in the form “If I change this, then that will happen, because of this reason.” It names the one thing you'll change and the measurable result you expect.
What are independent and dependent variables?
The independent variable is the one thing you change on purpose. The dependent variable is what you measure to see the effect. Everything else — temperature, amounts, timing, equipment — is controlled: kept the same so it can't explain your results.
How many variables should a science fair project have?
One independent variable, one dependent variable, and as many controlled variables as it takes to keep the test fair. Changing two things at once means you can't tell which one caused the result.
Is it okay if my hypothesis is wrong?
Yes — completely. A hypothesis the data rejects, with honest analysis of why, scores better than a vague guess that can't be proven wrong. Judges look for whether your conclusion follows your data, not whether you guessed right.
How many trials should I run?
At least three at every setting of your independent variable. Averages from repeated trials are what separate a result from a fluke, and judges ask about trial count directly.