Statistics · Ch 1 — Collection of Data
The Questionnaire — Design and Editing of Data
The Questionnaire — Design and Editing of Data
A questionnaire is a formally structured list of questions prepared in advance to collect specific information from respondents, used in several of the primary-data methods above (mailed questionnaire, schedule method, and often direct personal interviews too). Because the questionnaire is the single tool that stands between the investigator and the information needed, its design has a direct bearing on the quality of the data eventually collected.
Qualities of a good questionnaire
- Questions should be few, short, and simple — a long or complicated questionnaire discourages honest, complete responses.
- Questions should be arranged in a logical order, moving from simple/general to more specific.
- Questions must be unambiguous — each respondent should interpret a question the same way.
- Questions must avoid leading or suggestive wording that hints at a "correct" or expected answer.
- Personal, sensitive, or embarrassing questions should be avoided, or asked with care and reassurance of confidentiality.
- Wherever possible, questions should allow simple, direct answers (yes/no, a number, a tick-box choice) rather than open-ended essay responses, to make later tabulation easier.
- A pilot test on a small group before the main survey helps catch confusing or poorly worded questions in advance.
Once data is collected — whether through a questionnaire, a schedule, or any other method — it must be edited before it can be classified, tabulated, or analysed. Editing of data means scrutinising the collected data for completeness (no important item left blank), consistency (answers do not contradict one another), and accuracy (figures are within plausible/reasonable limits), and correcting or querying entries that fail these checks before the data moves on to the next stage of the statistical investigation (classification and presentation — the next chapter). …
A formally prepared, structured set of questions used to collect specific information from respondents, typically used with the mailed-questionnaire and schedule metho …
The process of scrutinising collected raw data for completeness, consistency, and accuracy, and correcting or querying doubtful entries, before the data is classified, t …
A small trial run of a questionnaire or schedule on a limited group before the main data collection, used to detect and fix ambiguous, confusing, or poorly …