Sample Selection Q&A

How can I ensure that my consultation results are valid?
Ideally results need to be free from bias (error caused by the process of selection of respondents, the wording of the questions, the context in which the consultation takes place, inconsistency of questionning) and results need to be representative of the group being consulted.
It is important to be clear and transparent, in your introduction, about the methodology you have used to consult with your respondents so that the reader can judge the value of your results.
You need to define clearly your target group. Usually a group is too large for the researcher to survey all its members so you need to choose a small sample to represent the group which reflects the characteristics of the larger group from which it is drawn.
You need to be willing to use your judgement and logic as there are no set rules.

Q. I have a large group and want every member to have an equal chance of being selected.
A. You need to use random sampling. This can be achieved by choosing with an unpredictable method e.g. sealed envelopes, flipping a coin, throwing a dice, drawing straws, spinning a roulette wheel etc
Q. I have a data base with a large number of names.
A. You could systematically elect every Nth name from the list of members. This method is also random, assuming the initial data base is random.
Q. I have to consult with young people and ensure that each age group is represented but one age group has very few members. Random sampling may mean that no young people from this age group are represented.
A. You could select a certain number of each age group. This stratified sampling is especially useful if a particular section within the larger group is under-represented and ensures that the under-represented group are included.
Q. I don`t have much time and cannot afford to travel far for my consultation.
A. Sometimes it is necessary to consider these other factors as well and to sample for convenience, whilst accepting that your results will be less accurate than a random sample. It is especially important to acknowledge these limitations in your report.
Q. I have to capture the views of a very large group which would be unmanageable.
A. You could choose a group to represent the large group. e.g. one class in a year group to represent that year group, as long as you are confident that the characteristics of the sample group represent the larger one.
Q. It is difficult and costly to identify respondents as they are scattered in a number of institutions over a wide area.
A. You could contact initial subjects and ask them to identify further subjects. This will produce less accurate results because it is less likely that the sample will be a cross-section. Again you need to acknowledge the limitations of this methodology in your report.