With the growth and availability of Online survey tools (like Survey Monkey) it may seem like a fairly simple task to conduct your own market research surveys. Just ask some questions and collect their responses and you have low cost insight! However there are many pitfalls and traps inexperienced researchers can fall into.

Follow these 10 tips in your survey design and ensure that your online research project is a success.

1. Set clear research objectives                                                                 

From the outset you should be clear about what you want to achieve. Your research objectives …specifically what you wish to find out…are your blueprint for designing a survey and if your objectives are clear, your questionnaire will follow suit. There is little point asking questions that do not relate to your research objectives. The key here is to be succinct and straight to the point with your aims, and don’t start by trying to word the questions (which is a common mistake!).

2. Think ahead to your findings

You should consider how you will capture and analyse the collected data before you start the process. A great tip is to test the survey on colleagues (a pilot survey), as it will allow you to download the data to see how it is structured (as well as highlight any flaws in the survey design). If you are looking to segment your findings make sure you allow for this by asking the right questions. Some survey software even has an option to create dummy data before a survey is live. All of these approaches can help you visualise your data and anticipate any issues. Think ahead to your findings.

3. Questions should be direct and clear

You may only have one chance to ask the respondent a question so ensure it makes sense and cannot be misinterpreted. Consider the following question: do you drive a car? This question is ambiguous and can be interpreted in many ways; personally owning a car, having access to a car in a household, having a company car or a car for business use, or even having passed your test. If you want to know if the respondent owns a car then ask them directly and unambiguously. Put yourself in the position of the respondent, ask yourself the question and try and work out how you would respond.  Better still ask colleagues to pilot and test your question for comprehension.

4. Avoid double-barrelled questions

A common mistake! Double-barrelled questions ask two questions and leave the respondent unsure about how best to respond. For example, ‘did you enjoy the play and was the seat comfortable?’ Ask one question at a time and if they are similar, consider making the font bold to ensure the respondent can easily see the difference.

5. Avoid leading questions 

Some questions appear to lead the respondent. For example, ‘do you agree with the view that…?’ suggests a particular response and will influence some respondents.

6. Answer options should be unambiguous and balanced

Ensure your answer options cannot be misinterpreted.You should always avoid terms such as ‘often’ and ‘regularly’ as measures of frequency. They can be ambiguous as respondents may operate with different frames of reference. For example, ‘regularly’ may mean monthly to one person whereas it may mean weekly to another. Whenever possible try to ask about actual frequency such as, ‘more than once a month’, ‘once a month’, ‘2 or 3 times a month’ etc. For closed questions, ensure the available answers are balanced e.g. extremely satisfied, satisfied, neutral, dissatisfied, extremely dissatisfied.

7. Avoid technical terms

If your language is confusing it may lead to a respondent becoming frustrated and even terminating the survey. So avoid jargon at all costs! Use simple, clear and plain language that is straight to the point. If you are struggling, the Plain English Campaign has a useful A-Z of alternate words.

8. Question Order

It is important that the order of questions does not influence your results. Therefore, questions have to be asked in a sequence that does not unnecessarily prompt or remind respondents prior to capturing their initial thoughts. Demographic and classification questions are best left to the end of the sequence as these can be the least interesting from the respondents perspective. If sensitive questions e.g. age, income and gender are asked too early in the survey, it can lead to early termination.

9. Use of space and design  

A survey is no different from any other marketing material; it should be visually appealing – perhaps brand aligned – and easy to read. Avoid cluttering questions together at it may confuse respondents.

10. Piloting

Piloting is imperative, you should always strive to test your survey on a small number of respondents – if this is not possible test it on your colleagues. People may not understand what you are asking and you may need to modify your questions to get what you want. If you don’t pilot test and work out those problems, the results from the final survey may be inadequate. Verifying the survey in this way, allows revising and editing questions before going live.

I hope this advice is useful. As always your comments are most welcome.

Jim

Check out other related posts in this series including Quantitative research – when and why to use it?

Jim Law owns Wild Heather Digital Research. WHDR uses the latest technology to deliver primary research insights to organisations. We offer online quantitative and online qualitative research services, including the development and use of panels and broad based consultation programmes in addition to online group discussions, depth interviews and online survey.