Census and Survey Operations Editing and Proofreading Services

A census is an enormous, expensive, once-a-decade attempt to ask everyone the same questions and get comparable answers. Everything rests on the wording. A question that means one thing to a homeowner in a city and another to someone in a multi-generational household in a remote community does not produce two answers to the same question — it produces a data series that looks consistent and is not, and the error is invisible in the results and permanent in the record.

We edit what statistical agencies and survey operations produce — census and survey questionnaires in all their modes, question wording and response category sets, instructions and definitions accompanying questions, respondent communications including pre-notification, reminder and non-response follow-up letters, enumerator and field interviewer manuals, training material and scripts for field staff, privacy and confidentiality assurances, help-line and web support content, cognitive testing protocols and reports, methodology documentation and quality indicators, data release and dissemination material, and public communication campaigns encouraging participation. Our editors work on the source wording that governs every mode and every translated version, and we check that a question means the same thing to every population it will reach.

Question wording is where a census is won or lost, and the failures are specific rather than general. Double-barrelled questions that ask two things and permit one answer; response categories that are not exhaustive, forcing people into "other" and destroying the variable; terms like "household", "usual residence" and "employed" that carry technical meanings respondents do not share; and questions about identity or income that produce refusal when the framing feels like an accusation. We work on these so each question asks exactly one thing, so the definition a respondent needs sits immediately before the question rather than in a glossary, so the categories are tested against the actual population rather than the one the agency imagines, and so sensitive questions state why the information is collected and what it is used for at the point of asking — which measurably reduces item non-response. Where cognitive testing has already shown a problem, we make the revision address what testing found rather than what the drafters preferred.

Everything you send is treated in strict confidence, including questionnaires in development and testing results. We are editors rather than statisticians or methodologists, and question design and its testing remain yours. What we can do is make the wording plain, unambiguous and clean at the source, which is where every mode and every translation inherits its clarity.