Surveys, questionnaires and modes of collection

A survey gathers information from individuals. The usual instrument is a questionnaire (filled in by the respondent) or an interview schedule (filled in by the interviewer).

A good questionnaire follows these rules:

• Keep it short: as few questions as possible.
• Make questions easy to understand and answer.
• Put them in a comfortable order, moving from general questions to specific ones.
• Make each question precise and clear, not ambiguous (open to two meanings).
• Avoid double negatives ('Don't you think it is not right…?').
• Do not ask leading questions that push the respondent towards one answer.
• Do not put alternatives inside the question itself.

Questions can be closed or open:

• Closed-ended questions give fixed choices: two-way (yes or no) or multiple choice. They are easy to score and analyse, but the respondent's true answer may not be among the choices, so an option like 'Any other' is often added.
• Open-ended questions let people answer in their own words: 'How can the problem of unemployment be solved?' They give individual answers, but are hard to score and compare.

Data can be collected in three ways:

• Personal interviews: the investigator meets each respondent. This gives the highest response rate and allows clarification of doubts, but it is costly, slow, and the interviewer may influence the answers.
• Mail questionnaires: the cheapest method, which can reach remote areas and has no interviewer influence. But many people do not reply, and respondents must be able to read and write.
• Telephone interviews: cheaper and quicker than meeting people, but not everyone has a phone.

Before the full survey, the questionnaire should be tried out on a small group. This pilot survey (pre-testing) shows up weak questions, unclear instructions, how well the enumerators (the people collecting the data) work, and the likely cost and time.