The Key Principles of Effective Questionnaire Design
Seven steps:
- Decide what information is required
what do you want to learn about - Make a rough listing of the questions
Any and all questions, narrow them down and rephrase later - Refine the question phrasing
Refine how the questions are worded - Develop the response format
Pre-coded list, open ended, scales? This is as important as the questions - Put the questions into the appropriate sequence
Logic and flow, easy questions then harder - Finalize the layout of the questionnaire
Format it and include good instructions - Pretest and revise
Here have people test it and find out what works and does not work. Test it out like it would be normally done (phone, interview, etc.)
Designing Effective Questionnaires
The purpose of any questionnaire is to extract data from respondents. There are five people to consider when designing a questionnaire:
- Client – wants answers to problems
- Researcher – needs to get the information while making sure that the interviewer can use the questionnaire, the questions are interesting to the respondent, and the questionnaire matches what the client needs.
- Interviewer – wants an easy to use questionnaire
- Respondent – wants to enjoy what he is doing as well as not feel they can not answer the questions as worded. They may also want a reward for doing the survey.
- Data-processor – wants a questionnaire that result in minimum error and efficiently
It is normal for a questionnaire to go through 7 or 8 revisions. If not they may be sloppily written.
Five sections in a questionnaire:
- Respondent’s identification data
- Introduction – scripted area asking for help and why
- Instructions – explain each section and how to navigate them
- Information – questions and responses
- Classification data – what demographic is the respondent in
Ten things to think about in designing a questionnaire
- What are the objectives of the survey?
You have to know what you need if you expect to get answers - How will the interview be carried out?
Know the method and design for it. - Think about the introduction
It has to be the same for all so that it cannot influence people but not have to sound wooden. Important because fail to engage the respondent and you loose the interview - Think about the formatting
Plan on how and when it will be given and plan for those elements. Leave space for comments. - Think about the questions from the point of view of the person answering them.
Do not get greedy for information and make the survey to long or ask impossible questions. - Think about the possible answers at the same time as thinking about the question
Understand where the possible problems could come if you do not allow the proper responses. - Think about the order of questions
Group in logical sequence - Think about the types of questions
open ended, closed, scales? - Think about how the data will be processed
Knowing how it will be tabulated will make question design easier - Think about interviewer instructions
Make them clear and easy to follow
12 things to look for when creating questions
- Questions should not be biased – do not lead the respondent to an answer
- Make questions as simple as possible
- Make questions specific – even of you have to break #2
- Avoid jargon and/or shorthand – do not assume the respondent knows them
- Steer clear of sophisticated or uncommon words
- Avoid ambiguous words - usually and frequently
- Avoid questions with a negative in them - ‘Do you never’ v. ‘Do you ever’
- Avoid hypothetical questions – they can never be accurate
- Avoid words that could be misheard – fifty v. fifteen
- Desensitize questions by using response bands - woman quicker to give age if it is a wide range of ages
- Ensure that fixed responses do not overlap – like the age bands above
- Allow for other in fixed response answers
Matching the questionnaire to the research objectives
Structured questionnaires are used for large studies. Anticipate all possible answers of the prompted questions so they can be coded.
Semi-structured questions are often used in Business to Business environments. They are generally look for a mix of qualitative and quantitative information.
Unstructured questions are used in qualitative research. Often survey person acts as a tour guide.
No matter what the type, long questionnaires and/or repetitive questions will kill the data collection. Vary the type of questions.
An Introduction to Different Question Types
Questions collect three types of information: behavior, attitudes, and classification.
Behavioral Questions
Designed to find out what people do (product a v. product b). Record facts not opinions.
Attitudinal Questions
Answers the questions of why people do the things that they do (buy product a because they are environmentally friendly). Key words are why, how, which, who, where, what. Questions with scales are often used here.
- Verbal scales – use word ranges from very likely to not likely at all
- Numerical rating scales – 0 low to 10 high, 5, 7, & 10 are popular choices.
- Use of adjectives – have the respondent give a word to describe what you are asking
- Positioning statement – agree strongly to disagree strongly
- Ranking questions – put items in list in order of preference
Classification questions
These look to make sure that the person interviewed helps to give a pool of the type of people they are looking for. Types of questions include: gender, marital status, socio-economic status, type of work, how many hours worked.
Key Terms in Questionnaire Design
Questionnaire – set of common questions laid out in a standard and logical form to find out a respondent’s attitude and behavior.
Question – framing of what you need to know in a carful format as to not lead the customer to give a unbiased response.
Open ended questions – questions that invite a free range of responses,
Closed questions – questions whose responses are in a preordained answer set
Direct questions – measures exactly what it appears to be asking
Indirect questions – disguises the true need of the question by asking something that will reveal what you want
Multiple response questions: allow for multiple choice to answer any particular questions
Prompted questions – gives responses for user to choose, closed questions are prompted.
Unprompted questions – respondent can give any answer, open ended questions are unprompted
Response codes – the answers to closed questions
Question grid – questions laid out in a way as to have a header with the responses and questions as rows, like a spreadsheet, done to save space
Rating scale – some visual representation from low to high that allows for respondent to put his choice on the scale
Routing – instructions that tell you where to move to if certain responses are made
Trade-off questions – gives the respondent certain amount of points to spread amongst the replies with the points given equal to the points assigned
The Respondent, the Interviewer and Questionnaire Design
A good questionnaire should seem like a conversation between the respondent and the interviewer. This is accomplished by:
- Easy to answer questions at start then harder at end
- Questions likely to to interest should be earlier on
- Questions should be in logical order
- Filter questions should be grouped together
- Helpful to have an introduction before each change of topic
- Personal, complicated questions should be near the end
Getting surveys done can be difficult since respondents see so many surveys and feel they are giving up time with little in return. The interviewer needs to engage the respondent as quickly as possible with no long introductions. He needs to also get the respondent to trust him with just his words and his voice. Though we use the questionnaire as a script it is a good idea for the interviewer to encourage the respondent as they go through so that they can get a complete interview.
Advertisement
 | Questionaires for Research. An Annotated Bibliography on Design, Construction and Use |
Posts
