A useful inquiry states what will be measured, where, when and for whom before data collection begins.
Start with a question that evidence can answer. A textbook, news item or website may make you curious about parks, but "Are parks good?" is too broad. Try: "Among residents of this neighbourhood, how does distance from the park relate to weekly visits?" State how you will measure distance and the period in which visits are counted.
A hypothesis is a prediction you can test. For example: "Residents living farther from the park report fewer visits in the previous seven days." This predicts a relationship without assuming distance is the only influence.
Primary data are collected for your investigation. Secondary data already exist, such as a published map or population table. Quantitative data are numbers, such as distance or visit counts. Qualitative data describe experiences and meanings, such as why a route feels difficult.
You can combine the two data types in either order. Survey distances and visits first, then use interviews to explain an unexpected result. Or start with interviews: if residents mention a difficult crossing, a later survey can test how many people face that barrier.
Choose an aim, study area, sample size and timetable that fit your team, equipment and access. Use safe, approved sites and avoid damaging habitats. Participation must be voluntary, and you should collect no unnecessary identifying details. Run a small trial, called a pilot, to find unclear questions or practical problems before collecting all the data.
Step by step
Make the question specific
State whom you are studying, where and when. Name the things you will measure, called variables.
Choose the evidence
Use numbers to find patterns and people's accounts to help explain them.
Try the method first
Use a pilot to check the wording, equipment, time, safety and consent arrangements.
Worked example: Improve the hypothesis
"People like parks" does not specify a test. "Residents farther from the park entrance report fewer visits in the previous seven days" identifies two variables, a population and a period. It predicts a relationship; it does not assume distance is the only cause.
Watch out for this
You must always collect numbers before interviews.
Either order can work. A survey can reveal a pattern for interviews to explain. Interviews can also suggest a question to test in a later survey.
Check your understanding
Which change makes a hypothesis more measurable?
- Replacing all variables with the word better.
- Defining visits as the number reported for the previous seven days.
- Deciding the result before collecting data.