2. Reasonable structure and strong logic. The arrangement of questions should have a certain logical order, which conforms to the respondent's thinking program. Generally speaking, it is easy before difficult, simple after complex, concrete after abstract.
3. It's easy to understand. The questionnaire should make the respondents clear at a glance and be willing to answer truthfully. The tone of the questionnaire should be friendly, in line with the understanding and cognitive ability of the respondents, and avoid using technical terms. Take some skills to investigate sensitive questions, so that the questionnaire can be answered reasonably, avoid subjectivity and suggestion, and avoid the distortion of answers.
4. Control the length of the questionnaire. The time to answer the questionnaire is controlled at about 20 minutes, which neither wastes a question nor misses a question in the questionnaire.
5. It is convenient for data inspection, collation and statistics.
The procedure of questionnaire design includes the following steps: [2]
Grasp the purpose and content
The first step of questionnaire design is to grasp the purpose and content of the survey, and the essence of this step is to clarify the information needed for questionnaire design. This is also the first step of scheme design. For researchers who are directly involved in the design of the research scheme, they can also skip this step and start from the second step of questionnaire design. However, for those researchers who have never participated in the scheme design, the first task when designing the questionnaire is to fully understand the purpose and content of this survey. Therefore, it is necessary to seriously discuss the purpose, theme and theoretical hypothesis of the research, read the research scheme carefully, consult and discuss with the scheme designer, and make the problem concrete, organized and operable, that is, turn it into a series of measurable variables or indicators.
data collection
Design is not a simple imagination. In order to design the questionnaire perfectly, researchers need to know more. Questionnaire design is a technology that needs experience and wisdom. It lacks theory, because there is no scientific principle to ensure the best or ideal questionnaire. Questionnaire design is an art, not a science. Although there are some rules to follow to avoid mistakes, good questionnaire design mainly comes from the creativity of skilled researchers [3]? .
There are three main purposes to collect relevant information: first, to help researchers deepen their understanding of the problems investigated; The second is to provide rich materials for problem design; The third is to form a clear concept of overall goal. Interviewing individual interviewees when collecting information is helpful to understand their experiences, habits, education level and understanding of questionnaire questions. We know very well that the problems that apply to college students may not necessarily apply to housewives. The greater the group difference of the respondents, the more difficult it is to design a questionnaire suitable for the whole group.
Determine the investigation method
Different types of survey methods have influence on questionnaire design. In face-to-face investigation, respondents can see the problems and talk face to face with the investigators, so they can ask long and complicated questions of various types. In the telephone interview, the respondents can talk to the investigators, but they can't see the questionnaire, which determines that they can only ask short questions. The mailing questionnaire is filled out by oneself, and there is no direct communication between the respondents and the researchers, so the questions should be simpler and detailed. In computer-aided access (CAPI and CATI), we can realize more complex jumping and random arrangement problems to reduce the deviation caused by sequences. Questionnaires for face-to-face interviews and telephone interviews should be designed in a dialogue way.
Determine the content
Once the type of access is determined, the next step is to determine the content of each question: what should each question include, what should the questionnaire ask, whether it is comprehensive and relevant.
Here, for each question, we should ask:
(1) Is this question necessary?
(2) Do you need a few questions or just one?
Our principle is that each question in the questionnaire should help to provide needed information or serve a specific purpose. If you can't get satisfactory usage data from a question, then the question should be cancelled.
Of course, sometimes, you can "deliberately" ask questions and answers that are not directly related to the information you need. For example, asking some neutral questions at the beginning of the questionnaire can make the respondents willing to intervene and establish friendly relations, especially when the topic of the questionnaire is sensitive or controversial. Sometimes, some questions are "filled in" to cover up the purpose of the investigation or the funded (or entrusted) unit of the project.
The first principle is to determine the necessity of a question, and then the second principle is to confirm the sufficiency of this question to the information obtained. Sometimes, in order to get the information you need clearly, you need to ask several questions at the same time.
For example, the question of "why" in the choice of goods and schemes mostly involves two aspects: (1) "Not good-looking, but comfortable", (2) "Not comfortable, but good-looking" and (3) "Not good-looking and uncomfortable".
Here, in order to obtain the required information, two different questions should be asked:
(1) "Do you think XXX brand clothes look good?"
(2) "Do you think XX brand clothes are comfortable to wear?"
When determining the content of each question, researchers should not assume that the respondent can provide accurate or reasonable answers to all questions, nor should they assume that he will be willing to answer every known question. Investigators should try their best to avoid questions that respondents are "unable to answer" or "unwilling to answer".
The occurrence of "unable to answer" may be that the respondents "don't know", "can't remember" or "can't express". In the case of "not knowing", ask some "filtering questions" before asking, that is, measure past experience and familiarity, so as to filter out those respondents who don't know the situation.
Secondly, respondents may not remember some of the survey contents.
The results show that the ability to recall an event is influenced by three factors: (1) the event itself; (2) the time span of the event; (3) Whether there are other events that may be helpful to memory. The questions reviewed in the questionnaire may be useless or beneficial. Useless memories generally lead to underestimation of the actual situation.
For example, ask the respondent to answer the question and answer the question without any hint:
"What products did you see on TV last week?"
This is an example of useless memory.
If you list the names of a series of products or enterprises, then ask:
"Which of the following enterprises or products did you see advertisements last week?"
This is a helpful memory, which can stimulate the memory of the respondent by giving some hints.
Respondents are unable to express their answers to some types of questions. For example, it is often difficult to express accurately when respondents ask what atmosphere they like to eat in restaurants. However, if some alternative answers are given to describe the hotel atmosphere, respondents can point out which one they like best. Otherwise, if they can't express themselves, they may ignore this question and refuse to answer the rest of the questionnaire. So we should provide some help, such as pictures, maps and descriptive words, to help them answer.
Respondents are "unwilling to answer" questions for several reasons. First, respondents have to make great efforts to provide information; Secondly, some questions in the survey are inconsistent with the background of the survey (for example, the consumption of ordinary goods and personal privacy are not consistent in the same questionnaire); Third, the reasonable purpose of the investigation, without which the respondents are unwilling to provide the investigation; The fourth is a sensitive issue.
There are several ways to encourage respondents to provide information they are unwilling to provide:
(1) Put sensitive questions at the end of the questionnaire. At this point, the respondents' alertness has been greatly weakened and they are willing to provide information.
(2) Add a "preface" to the question and answer, explaining the background and * * * situation of relevant questions (especially sensitive questions)-to overcome the respondent's fear that his behavior does not conform to social norms.
(3) Using "third party" technology to ask and answer questions, that is, to intervene in the question from the perspective of others.
Deciding structure
Generally speaking, there are two kinds of questions in the questionnaire: closed questions and open questions.
How to design a questionnaire survey
Open-ended questions, also known as unstructured questions and answers, are freely answered by the respondents in their own language, without providing specific questions for choosing answers. For example:
"Why do you like Nike TV commercials?"
"What do you think of the current state-owned enterprise system reform in China?"
Open-ended questions allow respondents to fully express their views and reasons, and they go deeper and sometimes get unexpected answers from researchers. Its disadvantages are: there is a lot of useless information in the collected materials, which is difficult to make statistical analysis, and the records of face-to-face interviews of investigators directly affect the survey results, and may be rejected because of troublesome answers.
Therefore, open questions are very helpful in exploratory investigation, but they do more harm than good in large-scale sampling investigation.
Closed question and answer, also known as structured question and answer, stipulates a set of alternative answers and a fixed answer format. For example:
What are the main factors you consider when choosing to buy a house?
price
(b) area
traffic
(4) Surrounding environment
design
(f) Building quality
(g) Other _ _ _ _ _ _ _ _ (please specify)
The advantages of the closed problem include the following aspects:
(1) The answer is standardized and easy to code and analyze.
(2) Respondents are easy to answer, which is conducive to improving the questionnaire recovery rate;
(3) The meaning of the question is clear. Because the answers provided help to understand the meaning of the question, it can prevent the respondents from refusing to answer because they don't understand the meaning of the question.
The closure problem also has some disadvantages:
(1) It is difficult to detect the respondents' incorrect understanding of the topic;
(2) There may be "order deviation" or "position deviation", that is, the respondents' choice of answers may be related to the arrangement position of the answers. Research shows that respondents tend to choose the first or last answer, especially the first answer, for declarative answers. For a set of numbers (quantity or price), tend to take the middle position. In order to reduce the order deviation, several forms of questionnaires can be prepared, and the answers of each form are arranged in different order.
other
Decide on the wording of the question
Arrange the order of questions
Determine the format and typesetting
Draft the first draft of questionnaire and pre-survey.
Make a formal questionnaire
fold