Monday, October 19, 2015

Marketing Research 101: Six Commons Mistakes in Survey Questionnaire Design

Online surveys have become the cornerstone of the market research industry. They quickly and easily allow businesses to gather consumer data, which they can use to: enrich products, alter marketing campaigns, and tailor messaging. Unfortunately, modern technology, which has improved the ease with which companies can generate surveys and analyze results, has also spawned an era where an increasing number of surveys are poorly formulated, limiting responses or skewing data to misrepresent customer intentions. By applying a few basic tenants of survey design, we easily increase engagement and improve data accuracy and overall quality.

Questions that Affect the Answer
1.       Leading Questions: Questions should be phrased in a way that suggests all responses are equally viable
2.       Loaded Questions: Write questions that do not inherently encourage the participant to be less-than-straightforward with the response

Questions that Result in Unclear Data
3.       Double-Barreled Questions: Questions that ask two things and offer only one opportunity to answer do not provide usable data
4.       Questions with Absolutes: Unless the subject is clear-cut (i.e., ‘Are you a man?’), the question should provide enough choices to represent participant response

Questions that Confuse
5.       Jargon: Unless the participant pool is very specific, such a ward of doctors at a local hospital, questions should avoid using abbreviations, industry-specific language, or slang. Using clear, common language insures that 100% of the participant pool has an equal understanding of the content
6.       Offering Too Many Options: Potential responses to online survey questions should be limited to five or six. Offering more options can be confusing to the participant, clutters the survey, and may negatively affect survey completion

While there is much more to online survey design than adhering to these six principles, avoiding the most common errors of question formulation will greatly increase a survey’s return rate, accuracy, and relevancy.

About the Author:  As Global Marketing Associate, Tara Wildt leads Lightspeed GMI’s digital and interactive marketing platform, branding initiatives and social media strategy. In her role, she develops creative solutions and collateral for Lightspeed GMI’s product suite and plays a key role in the company’s strategic development. In addition, she has oversight of the company’s philanthropic and community outreach initiatives.  


Horst Feldhaeuser said...

Hi Tara, nice reminder of some basic rules. Would have been good to add a few examples (good & bad) for the less experienced survey writers. I.e. what does a leading or loading question look like.

Dr. Gordon C. Bruner II said...

Thanks for beginning this discussion, Amanda. Let me add that far too little thought is given to validating measures. I wonder how many survey developers even know what that means. It seems to me as if many developers believe that if they think a question (or set of questions) is "good" then there is no need to test that assumption. Maybe that is okay for some "minor" things being asked in a survey but for others, such as measuring mission critical issues such as purchase intention, satisfaction, and loyalty, validation is critical. (BTW, HOW to validate survey measures is a whole other topic worthy of discussion.)