Face-to-face interviews (perhaps supplemented by observation and taking measurements) are likely to be the only method for the collection of survey data that is suitable for use in populations with low levels of literacy. Verbal questioning can by supplemented by use of visual aids. Ideally, interviews should be conducted in private by an interviewer who is unknown to the respondent. Face-to-face interviewing is likely to generate better quality data than other forms of interview or questionnaires in long, complex surveys but represents an expensive approach to data collection.

Telephone interviewing is relatively cheap and has been the method used by much market research in more developed countries. It usually gives similar results to face-to-face methods, but leads to under-representation of those without landline telephones (or who are rarely at home). Refusals (including the use of answering machines to screen calls) can limit the representativeness of telephone-based surveys. Refusals have become an increasingly severe problem since the spread of the practice of conducting phony surveys in order to cold-sell over the telephone.

Using self-administered questionnaires can reduce the costs of collecting survey data. However, the response rate tends to be low if the questionnaires are distributed by post. It is important to keep self-administered questionnaires short and simple in order to maximise the response rate and the accuracy of the responses. A risk also exists that self-administered questionnaires will be completed by proxy respondents instead of the person who was intended to do so and that the accuracy of the answers will suffer as a result.

Self-administered questionnaires can be the best way of collecting sensitive information that respondents may not want to share in a face-to-face interview but are prepared to report anonymously.