How response alternatives affect different kinds of behavioural frequency questions

Responses to behavioural frequency questions in surveys and questionnaires form the bases for much social and behavioural research. The choice of response alternatives for these questions can affect responses. By embedding questions in national surveys, we demonstrate that the way in which response alternatives affect responses depends on whether the question uses a vague, an ambiguous or a relatively well-defined phrase to denote the target behaviour. We explain these results and discuss the implications of our findings for survey research.