Asking questions and influencing answers.

Abstract

On the foc-e of it, suJVey interviews are simple. An interviewer steps into the home of a randomly selected member of the public, asks a series of questions, records the anllwers, and departs with new facts or opinions to add to her (.'Oile(.1ion. (For (.'Onvenien(.-e let us think of the interviewer as female and the respondent as male.) The information she takes away is determined by the questions she asks-how they are worded and what they require of the respondent. Properly designed, they will give her the facts and opinions she w.mts. But this view of suJVey interviews is too simple, as the history of suJVeys has shown again and again. How a question is worded makes a differen(.-e, but so do many other factors-how the question is introduced, what questions come before and after, what answers are allowed, and much, much more. The factors are so diverse that they may seem impossible to account for. Even wording is mystifying. From the perspective of language use, many of these factors aren't so mysterious. At least this is what we will argue. What makes them seem that way is the common misconception that language use has primarily to do with words and what they mean. It doesn't. It has primarily to do with people and what they mean. It is essentially about speakers' intentions-what speakers intend in choosing the words they do, and what their addressees take them as intending. Once we understand the role of speakers' intentions in language use, we will find many of the problems of suJVey design more tractable. Our goal is to convince you that you can't understand what happens in survey interviews without understanding the role of intentions in language use. We will begin by describing five basic principles of language use that

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 210 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.