Using the Citizens’ Rail survey as an example:

Who are the target audience?
In this case, people who do not currently travel by rail, located in the five catchment areas: Aachen, Devon, Lancashire, Heerlen, Pays-de-la-Loire

What is the scale (geographical or otherwise) of the population?
In each catchment area, the population numbers served as a guide for how many survey questionnaire responses would be needed for the sample to be ‘representative’

How do you reach the numbers of people you need?
How do you reach the numbers of people you need? A choice must be made in terms of your available resources and how they can be best employed. For the Citizens’ Rail survey, a flyer was used with a weblink to the survey questionnaire. This method was used because the scale was too great to send out people to conduct the surveys, without costing a lot of money and time.

Where do you sample to get the most representative data?
Where do you sample to get the most representative data? For the Citizens’ Rail survey, flyers were delivered to every 5th house on every street in the area. Defining the boundaries of an area is important so that you know where to conduct the survey. If the survey had been interviewer administered, a sensible choice of location would be a busy city centre.

When are you going to collect data in order to ensure your sample accuracy is maximised?
For Citizens’ Rail, it was important to ensure data was collected before the ‘intervention’ (the introduction of new station facilities) was completed. Time of day is an important factor in sampling, as research should capture the broadest range of people.

This is one example. For different research methods the sampling strategy is different. For example, interviews typically only need 30 or less respondents. The researcher therefore has to consider how to access the most suitable people to contact for interviews.