Stories. Questions. Different perspectives.
Browse WAURE
← Advice & Questions
Education · Advice & Questions ·

Should we change our hypothesis after the experiment did not show the pattern we predicted?

Education category photographEducation

“We could just write that we expected this,” my partner said, looking at the results sheet.

For our school science project, we predicted that changing the folds in a paper bridge would produce a particular pattern in the load it held. Our trials didn't show that pattern clearly. The original prediction is in our planning notes, but the final report hasn't been written.

I don't want to alter the observations. I also don't want to rewrite the prediction as though we knew the outcome before testing. My partner thinks the project will look unsuccessful if the hypothesis isn't supported.

Science Buddies' conclusion guidance recommends explaining whether results support the original hypothesis and discussing limitations and further work.

How should we write this up honestly and still show what we learned? We can examine the method and suggest another test, but I don't want an uncertain result turned into either a fake success or proof that the question was worthless.

Discussion

5 community replies

Keep the original prediction identified as the original. State what the trials showed and how that compares with it. A later explanation or revised hypothesis can be labelled as something developed after seeing the results.

Ask the teacher how the project is assessed rather than assume an unsupported prediction automatically loses the important marks. You can demonstrate careful work without claiming a pattern you did not observe.

Review the method for specific limitations you can support. Do not invent an equipment problem simply because the result was disappointing. Equally, an unclear pattern in these trials does not establish that the tested change could never matter.

Separate the report's conclusion from the next experiment you would propose. Explain what uncertainty that follow-up would address, rather than recommend repeating until the expected answer appears.

The follow-up needs its own honest starting point. Record the new prediction before testing, if you go ahead, and keep the first set of results. Then a changed plan becomes part of the investigation rather than a replacement history of it.

Join the discussion

More in Education