Questions are the heart of marketing research, but they are often not given the thoughtful consideration that they deserve in that role. Especially when companies choose to do their marketing research in-house, poorly written questions can lead to seriously bad data.
For example, let’s look at this simple question:
How many cars do you own?
Simple, right? Not so fast. There are three potential problems with this “simple” question:
- Who is “you”? Is it you, personally, the person taking the survey? Or is everyone in the household included?
- What do you mean by “own”? Is it the person whose name is on the title? The primary driver of the car? And what about leased vehicles? Are they included in ownership?
- What’s a “car”? Are trucks cars? In some communities, golf carts are a preferred form of transportation. What about vans? Are they cars?
As you can see, in marketing research nothing is ever really “simple.” A survey is like speaking to a room full of 400 individuals. You need to make sure that everyone understands each question in the same way, to the same level of thoughtful engagement, so that they can deliver high-quality responses. This is a tall order.
Questions are where the art and science of research meet. There is no “right” way to ask any question, but there are tips to help you produce better surveys full of excellent questions:
- Unambiguous wording. Like the example above, don’t just dash off your questions and assume all your respondents will know what you mean. Think about variations in your population’s demographics, lifestyles, geographies, and anything else that could lead them to interpret your question differently than you meant. While it is important to sound friendly and approachable, don’t push that so far that you can’t interpret your resulting data!
- Limit the Grids. Grid questions work well for quickly collecting comparable data. However, too many of them in a questionnaire is tedious, especially for respondents who may be taking the survey on their phones. Respondents will stop reading the grids and will “straight-line” their answers, so if you must use grids, keep them short and sweet.
- Stay neutral. Avoid loaded, leading, and biased question wording if you want good data quality. If you want some examples of biased, loaded and leading questions, look no further than political polls. (Or check out this article by Surveytown.) And, please, no double negatives! Questions using double negatives are misleading at best and will provide low-quality
- Limit response choices. Yes, you can probably list many, many responses to any question you can think of. But most of the time in marketing research, you don’t need to include every possible response. Include the most frequent responses and cover the rest by using an “other, specify” response. You can look at the verbatim responses after the data is collected and determine if there is another response category you want to add back for analysis.
- Respect your respondents. Willing, high-quality respondents are increasingly becoming an extremely scarce commodity. So, let’s treat them with the utmost respect. If you tell them it will take 10 minutes to complete the survey, take 10 minutes and not a second more. Use a completion scale to help respondents understand how much longer they have until they are finished with the survey. And, please, let respondents know you appreciate their participation by giving them a big “Thank You!” at the end of your survey.
- In today’s time-compressed, “I need it yesterday” world, pretesting has become a lost practice.Typically, we look at the initial results to make sure that the survey logic is working, and that the data looks viable. But a true pre-test with qualitative questions about what the questions mean, how the respondent answered, and especially whether the respondent was uncertain or confused about how to answer can drive critical question improvements that deliver better data quality. It will add a little time and cost to your project, but it will also ensure you get the results you need.
As marketing researchers, we spend much time thinking about sample source, sample size, methodologies, and analysis. Let’s add question design back to that list. One of the key reasons potential respondents break off surveys (or don’t even start them) is poor question and questionnaire design. Without good questions, you could be driving away respondents as well as wasting your own valuable time and money!
Need help with your questionnaire design? Contact Us to get started!