BIASES

Who is surveyed and how they are surveyed can be critical to the outcome of the research. Key issues include:

Sample size – were enough people asked? A bigger sample offers greater insurance against unrepresentative results. The number of people surveyed determines whether the results can be considered ‘statistically valid’. A caveat will then be added emphasising the degree to which any percentages might vary e.g. +/- 3%

Composition – do the people asked reflect the range of people whose experience you want to understand? For example, does the survey properly capture the views of relevant demographic groups? If surveying a specific group, does the sample include the full range of people whose views you are seeking?

Conduct – has the survey been conducted according to Best Practice and in the optimum circumstances? Again, a reputable company will use trained researchers to optimise the circumstances in which responses are obtained.