Principles of ordinary conversation suggest that respondents will tend to interpret two similarly worded questions in an interview as having the same meaning.

Satisficing may explain why some respondents tend to give “no opinion” or “don’t know” responses.

Closed-ended questions are more valid than open-ended questions because they offer standard response options.

Survey questions always should explicitly provide the “don’t know” response option.

A good source of survey questions is questions developed by other researchers.

Sensitive questions are best placed at the beginning when the respondent’s interest is high.

Providing a reference period such as yesterday or last week generates less accurate reporting of behavior than just asking respondents what they usually do.

One way of avoiding response bias tendencies is to vary the arrangement of questions and the manner in which they are asked.

Cognitive interviewing attempts to understand respondents’ thought processes in answering survey questions.

Behavioral coding and respondent debriefings are designed to make field pretesting more systematic and reliable.