1. The explanation of the questionnaire should be simple and clear. The interpretation of touching questionnaire is related to the quality and effect of investigation. Generally, it is necessary to explain the purpose and significance of the survey in a euphemistic and touching tone, especially to make the respondents feel that the survey is useful and meaningful to themselves, or to help others and inspire their enthusiasm for helping others. 2. Avoid using inaccurate words. Some adverbs and adjectives, such as "for a long time", "often" and "some", have different understandings and should be avoided or reduced as much as possible when designing the questionnaire. For example, "Do you often get sick?" "How many times did you get sick last month?" Or "How many times have you been ill in the past six months?" . "Where were you born?" "Where were you born (please indicate the province, city or 3. Avoid asking questions such as "How many cigarettes do you smoke a day? "This kind of question is conclusive. If the respondents don't smoke at all, they won't be able to answer. The correct way to deal with this problem is to add a "filtering" question before it. For example, "Do you smoke? ".If the answer is" Yes ",you can continue to ask questions, otherwise you can stop asking questions. 4. Avoid leading questions. Guiding questions mean that the questions raised imply the researchers' views and opinions, and the respondents may answer with this tendency. For example, "Some people think that passive smoking can cause lung cancer. Do you agree? 0= disagree 2= don't know 3= agree that passive smoking can lead to lung cancer. 0= disagree 2= don't know 3= agree 5. Sensitive Issues to Avoid Embarrassment and Taboos Respondents include taboo issues in local customs and national habits, issues involving personal interests, and personal privacy issues. For example, "Do you have extramarital sex? How many times? Investigation and handling of sensitive issues: 1. Interpretation: that is, write a paragraph before the question to eliminate concerns, or state in the introduction of the questionnaire that the respondents will be kept strictly confidential and explain the confidentiality measures to be taken. 2. Hypothetical method: Take a hypothetical conditional sentence as the premise, and then ask the respondent's opinion. 3. Transfer method: transfer the questions that should have been answered by the interviewee according to his own actual situation to the interviewee's explanation of his own ideas according to others' situation. 6. Avoid asking general, abstract or inaccurate questions. Concepts that are easily misunderstood should be clearly defined. For example, age is nominal and real; Income only refers to wages, or includes bonuses, subsidies, other income, and in-kind payment income; Family population consists of resident population and living expenses. 7. Avoid multiple questions. It is best to ask only one point for a project. If a project contains too many questions, it will make the respondents unable to answer them, which will bring difficulties to statistical processing. For example, "Are your parents intellectuals? What's your father's educational level? What's your mother's educational level? 8. Each question in the table has a clear purpose. For each question, it should be clear why this question is raised, what kind of analysis will this information be used for, and how to code and analyze it. 9. Ensuring the expression and order of information quality questions to the greatest extent is conducive to enlightening respondents, making questions interesting and easy to recall, and avoiding those questions that are difficult to answer, time-consuming and embarrassing. 10. Be concise, ask questions fluently, and pay attention to communication skills with respondents with the lowest education level. When one question turns to another, pay attention to the logical relationship, words and tone, such as from general to individual, from easy to difficult, etc.