Choosing a method
| English | Chinese | Pinyin |
|---|---|---|
| mixed methods | 混合方法 | hùn hé fāng fǎ |
| validity | 效度 | xiào dù |
| reliability | 信度 | xìn dù |
| research ethics | 研究伦理 | yán jiū lún lǐ |
Numbers and reasons answer different questions
- Quantitative methods measure how much and how many. Qualitative methods explain how and why.
- Mixed methods 混合方法 use both, and most good Level III projects do.
- The design question is not which is better. It is which parts of your question need which.
Validity and reliability
- Validity 效度 asks whether you measured what you claimed to measure.
- Reliability 信度 asks whether the measurement would repeat.
- A study can be reliable and invalid: a badly worded question returns the same wrong answer every time, consistently.
- Both must be argued for in the proposal, not asserted.
A badly worded question returns the same wrong answer every time. What is it?
It repeats consistently, so it is reliable — and it is not measuring what was claimed, so it is invalid.
Research ethics
- Research ethics 研究伦理 are marked: informed consent, anonymity, the right to withdraw, and no harm.
- Consent must be informed, which means participants know what the study is about before agreeing.
- ⚠ A proposal without an ethics section is incomplete, whatever the design is worth.
A proposal with an excellent design but no ethics section is complete.
Ethics are marked. Informed consent, anonymity, the right to withdraw, and no harm.
Food waste in a school canteen.
Quantitative: weigh what is thrown away, by day and by dish, for two weeks.
Qualitative: interview ten students about why they left food.
The weights say which dishes and which days. The interviews say why — portion size, a short lunch break, a dish nobody chose.
Neither alone supports a recommendation. Together they do, and the recommendation is what the report exists to make.
Weighing food waste tells you which dishes. What do the interviews add?
Neither alone supports a recommendation. Together they do, and the recommendation is the point.
Name one part of a food-waste study that needs interviews rather than weighing.
Example: "Why students left food — whether the portion was too big or the lunch break too short — cannot be weighed."
Choose the method after the question, never before. A student who has decided to run a survey and then looks for something to survey has reversed the only step that matters, and it shows in every section that follows.
Put the design steps in the only order that works.
Deciding to run a survey and then looking for something to survey reverses the only step that matters.
Do not collect data you have no plan to analyse. Every extra question on a survey lowers the response rate and adds a column nobody will use — and an examiner reads a bloated instrument as evidence the question was never sharp.