Brainstorm lists of possible project ideas. Choose a
project that interests you and that you can successfully investigate. The
best science projects involve a controlled experiment. Browse through the
ideas on this site or brainstorm your own list. List your ideas in the
form of a question.
Design an experiment to test your hypothesis. Each
experiment can only test one variable. For Instance, you could test the
effect of temperature on seed germination. The change of temperature is
your variable. You could not test two variables in the same experiment.
For example, you would not test the effect of temperature and moisture on seed
germination at the same time, because you could not determine which variable
caused the results you observed.
Each experiment needs a control. The control is an
additional trial done where nothing is changed. The control allows you
to decide whether the results were caused by the changes, or perhaps some
other factor. You run your experiment changing one variable at a
time, but you also run one trial each time where nothing is changed.
More than one trial should be done for each variable in
an experiment. For instance, how could you tell whether the seed not
germinating was due to the temperature, or due to a dead seed. If you do
the experiment 5 times and the seed never germinates, you could conclude that
it was due to the temperature. However, if 4 out of 5 seeds germinated,
you would know that it was a dead seed.