The (re-occurring) problem of questionnaire design
Questionnaires are key to a successful project, but we continue to see poorly designed questionnaires undermine projects...
A research project will contain many different stages, and each one is important to a succesful project!
At SwissPeaks we abide by our “6-week plan” which sets out in detail for clients the individual stages that need to be completed before a research project can enter the field. However, again and again we find that projects are being regularly undermined by poor questionnaire design. Basically, no matter how well the other stages of a project go, a poor questionnaire will cause problems throughout…from start to finish!
Consequently, this blog will seek to address some of the common problems we come across which all researchers should adhere to…
The top five things all researchers must do during questionnaire design
1. Ensure Consistency
You must be consistant in your design! A questionnaire should be like a report for a client – consistent throughout in layout and formatting. When building a report for a client you don't use different font types, you use structured control of font sizes, of colour, of layout and styles etc...
So why is this different for a questionnaire? It shouldn't be. Use a consistent layout theme. It's always evident to us when 2 or more people have developed a questionnaire together as each have their own styles, quirks, mannerisms. Make sure it is consistent. Using a set company brand and identity should also be used, alongside the use of "Tables" in MS-Word.
2. Make use of numbers
Numbers are important!When designing a questionnaire, it is important to assign a code value system (a number) to the answer categories.
This helps to assist the scripter in understanding the flow of that specific question, but more importantly, it will help with filter logic for any questions that are dependent on that question.
3. Use Interviewer Instructions
Every question must include interviewer instructions! These instructions are there to remove all potential confusion and ambiguity for the interviewer in the field, and therefore are key and must be used.
- READ OUT ANSWERS" and "DO NOT READ OUT ANSWERS" are commonly the default instruction that we use.
4. Specify question type
All questions should have the question type clearly defined!
It is no use having a nicely designed questionnaire if the respondent is not told what type of question they are answering. Therefore, during the design stage the four question types we recommend using are "MULTI CODE", "SINGLE CODE", "NUMERIC" and "VERBATIM".
These are useful for the interviewer, scripter and respondent, pending on data collection platform. They help to inform how an individual should be responding to a question. Without such information, it is likely a respondent will answer questions incorrectly.
5. Create logical Filter instructions
Make sure every question (or question set) is clearly defined by a filter instruction! - Please note, a "filter" is an "inbound" logic instruction.
When it comes to scripting the questionnaire, the scripter will need this level of detail to avoid making mistakes. In the filter instruction clearly state the filter definition in ENGLISH and in SYNTAX. E.g. IF WATCHED TELEVISION IN THE LAST 7 DAYS" is the English caption and "CODES 1 OR 2 AT B1" is the Syntax caption.
We believe questionnaire design to be crucial to any succesful project and so we hope you have found this useful! - Any questions on questionnaire design please do get in touch at Alex@swisspeaks.com