Coaching Centre
Choosing a research question
The kinds of investigation you can do, how to find a topic that suits you, how to narrow a broad idea into a focused, testable research question, and how to plan the project with a timeline and a logbook.
In 2015 a US company bought the rights to an old drug, Daraprim, used to treat a dangerous parasite infection, and raised the price from US$13.50 to US$750 a tablet. In 2016 a group of Year 11 students at Sydney Grammar School, working with University of Sydney chemists and the Open Source Malaria project, made the drug's active ingredient in their school lab. Their question was simple: can a school lab make it cheaply? Most good research starts with a question like that: specific, testable and something you actually care about.
What counts as a research project?
A student research project is an investigation you plan and carry out, from the question to the final report. It doesn't have to be a lab experiment. Scientists work in several ways, and so can you:
| Type | What you do | Example |
|---|---|---|
| Controlled experiment | Change one variable, measure another, keep the rest the same | How does the type of mulch affect water lost from potted soil? |
| Field study | Measure things where they happen, without changing them | How does the number of macroinvertebrates change along a creek below a stormwater drain? |
| Secondary data investigation | Analyse large data sets collected by others | Has the number of days over 35 °C in Penrith changed since 1970? (Bureau of Meteorology data) |
| Model or simulation | Build a model and test how it behaves | How does the angle of a solar panel affect its output across a year? |
| Design and build | Design something to solve a problem and test it against criteria | Which design of low-cost water filter removes the most cloudiness from dam water? |
Whatever the type, the steps are the same: question, research, plan, collect data, analyse, evaluate and report. This chapter follows those steps in order.
Finding a topic
Good ideas usually come from something you have noticed. Why does the steel drink bottle keep water cold longer than the plastic one? Does the cricket ball really swing more on humid days? Why do the tomatoes on one side of the garden do better? You can also look at science competitions, news stories about research, or the "further research" section at the end of a scientific paper.
Before you fall in love with an idea, check that it passes these tests:
- Testable: it can be answered with evidence, not opinion. "Which ice cream tastes best?" is a preference, not a science question.
- Measurable: you can record numbers (quantitative data), not just descriptions.
- Safe and ethical: no dangerous chemicals or equipment you can't control, no harm to people or animals (see 8D and 8E).
- Doable: you can get the equipment, and it fits in the time you have, including time to repeat things that go wrong.
- Deep enough: there is more than one thing to test, so you can extend it if the first experiment works.
From a topic to a research question
A topic like "water in the garden" is far too broad to test. A good research question names one independent variable (what you will change), one dependent variable (what you will measure, with units) and enough detail to show the scope. A useful pattern is:
How does [independent variable] affect [dependent variable] in [specific situation]?
Priya lives in Western Sydney and wonders whether mulch is worth buying. Turn her idea into a research question.
- Start with "Mulch and soil water". Choose the weakest option in each column and read why each fails.
- Now build the best question. Which check fails if you pick "the weather" as the variable to change?
- In "Sunscreen and UV", which choice is ruled out for ethical reasons, not scientific ones?
- Try "Paper towels". What kind of graph will the best question need, and why?
Working to a timeline
Most schools give four to ten weeks for the project. The single biggest reason projects fail is running out of time: a plant dies, the weather turns, a sibling waters your control pots. Build a timeline backwards from the due date, and leave a spare week for disasters.
Two things are often forgotten: a pilot test (a quick trial run that shows whether your method works before you collect real data) and time to get your plan and risk assessment approved by your teacher before you start.
Keeping a logbook
A logbook is the honest diary of your project, from your first ideas to your final draft. Professional scientists keep them too, and in research labs they can be legal evidence of who discovered what and when. Your teacher will use it to see how you worked, not just what you finished with.
- Use a bound book (not loose pages), number the pages, and date every entry.
- Write in pen. Cross out mistakes with a single line instead of using correction fluid, so nothing looks altered.
- Record raw data straight into the logbook, with units, before any calculations.
- Note conditions (weather, time), problems, changes to the method and why you made them.
- Glue in photos, printouts and sketches. Dot points are fine; it doesn't need to be neat.
Some schools allow a digital logbook (a shared document or an app). The same rules apply: dated entries, and never delete an entry, only add a correction.
Test yourself
Which is the best research question for a Year 10 project?
It names one independent variable, one measurable dependent variable with units, and the scope.
A student analyses 50 years of Bureau of Meteorology rainfall data for Dubbo. This is
The data was collected by others; the student's job is to analyse it.
Why should mistakes in a logbook be crossed out with a single line rather than erased?
A logbook is evidence of what happened, including errors.
What is the main purpose of a pilot test?
A quick trial run shows up problems while there is still time to fix them.
Label the features of a good logbook page.
Sort these questions.
A quick trial run of a method is called a test. The variable you change on purpose is the variable. Data recorded as numbers is called data. A dated record of the whole project is kept in a . A chart showing tasks as bars across the weeks is a Gantt .
Jordan wants to investigate "whether exercise is good for you". Explain why this is not yet a suitable research question, and rewrite it as a suitable question for a Year 10 project.
"Is exercise good for you?" is far too broad: "good for you" could mean fitness, mood, weight or sleep, and none of these is defined or measured. There is no clear independent variable (what kind of exercise, how much?) or dependent variable, and a project that pushes people to exercise hard raises safety and ethical issues. A better question is: "How does the time spent skipping (0, 1, 2 and 3 minutes) affect the heart rate of Year 10 volunteers, measured in beats per minute, one minute after stopping?" It has one independent variable, one measurable dependent variable with units, a realistic scope, and it can be done safely with volunteers who give consent.