Choosing a method, and the ethics check

Match the question to the method

• Survey: what many people think, do or report. Cheap and fast, but people misremember and like to look good.
• Experiment: whether one thing causes another. You change one variable and hold the rest still.
• Interviews: why people do what they do. Small numbers, rich detail, hard to generalise.
• Secondary data: questions someone already collected data for. Statistics Canada, open government portals, published datasets.
• Computation: patterns in data too big to read by hand. Needs some code, often less than you think.
• Simulation: systems you cannot touch directly. A model is only as good as its assumptions; write them down.

Before you collect anything, write down

1. Exactly what I will measure, and with what:
2. How many people or samples, and how I will pick them:
3. What result would prove my idea wrong:
4. What I will do if I get a result I did not expect:

The ethics check · every box, every time

• Everyone taking part knows what the study is and agrees, in writing or on a form.
• Anyone under 18 also has a parent or guardian's permission if your school or the study requires it.
• People can stop at any time, and can skip any question, with no consequence.
• Names are replaced with codes (P1, P2…) before anything is shared or written up.
• Nothing in the study could physically hurt, upset or embarrass someone.
• You are not studying people who cannot say no to you (younger siblings, people you supervise).
• A teacher or mentor has read your plan before you begin.

Honesty rules

• Report the results you got, not the ones you hoped for.
• Keep every raw data file, even the messy ones.
• If you drop data, say what and why.
• Credit every idea, dataset and figure you did not make yourself.