Improve and Conclude a PSLE Science Experiment

"Suggest one way to improve this experiment." "State the conclusion that can be drawn." The last parts of an experiment question look like a pile of different question types. They are not, once you sort them — and by this episode, you already know most of the pieces.

The idea in plain words

Asked to improve a set-up? Work in two steps, in this order. First find the flaw the question really describes. Then pick the fix that matches it. The matching fix, not the first one that comes to mind.

Common improvements include repeating measurements to check consistency, averaging to reduce random variation, and choosing a tool with a suitable scale. Each helps with a particular problem; none removes every source of error. Episode 6 explains these examples.

The third flaw is the only new one, and it is the one most often missed — so it's worth an extra moment. A variable that should have been kept the same was allowed to change. No amount of repeating or re-measuring fixes that, because nothing was wrong with the measuring — the second difference between the set-ups was there all along. The fix is fairness: keep that variable the same in every set-up.

A conclusion answers the aim you were set, and it has one boundary: it must stay inside the data you were given. Which way the result went — you can already do that from Episode 5. What is new here is just the boundary: don't say anything your readings did not show.

A prediction uses the results and relevant science to suggest what may happen at a value not tested. Keep other conditions the same, and check whether the trend is likely to continue. A prediction is not an observed result.

What a strong answer does

• Picks the fix the flaw actually calls for. "Repeat and average" is the fix everyone thinks of first, so it gets used where a different fix was needed. Was a variable allowed to change? Then the fix is fairness, not repeating.
• Claims only what the readings show. "Repellent keeps all mosquitoes away" goes past the data — the landings fell to 3, never to zero. Name the range you tested and stay inside it.
• Checks the basis for a prediction. Use the pattern and any scientific limits given in the question. Do not assume a line continues forever. Changing the cloth or the observation time may also change the result.

You've now built every part of an experiment answer, from naming the variables to writing the conclusion. If you want to try it on real questions, MCQ.SG marks open-ended questions point by point and gives feedback on every answer.

For more tips on improving your answers—from avoiding common mistakes to using precise science language—explore the complete PSLE Science OEQ Guide.