Online Surveys That Deliver

By: Kristin Bright

As the percentage of consumers with computers and broadband continues to increase, online surveys have become a reliable methodology for collecting data for select target groups. Following are tips for writing online surveys that will deliver results you can take action on.

    1. Online surveys must be brief. Keep it short! Nothing more than 10-12 minutes. Remember, a web survey is a self administered survey. Respondents tend to scan questions and get bored with too much content. Drop out rates increase markedly after 12 minutes.

 

    1. Online surveys must be clear and concise. Be wary of lengthy descriptions or explanations. Descriptions of product concepts should be written as you would in an advertisement - clear and concise. Lengthy descriptions will get lost in translation.

 

    1. Remember your audience. Be sure to consider your audience when writing all survey questions. Consider the age, income, education, and interests of your respondents when writing the survey questions.

 

    1. Give online survey respondents an "out." Respondents tire easily and become very frustrated if they are forced to answer questions that don't apply to them or they aren't given an option to choose "don't know." Giving respondents all possible options will help reduce a survey's drop-out rate.

 

    1. Use open-ended questions judiciously. Provide choice responses versus open-ended questions when possible. Some respondents who will not type anything if left on their own, but will select a response if given choices. Remember, they are no probes in online questionnaires. A good rule of thumb is that a respondent can answer 3 closed-ended questions in the time it takes to answer one open-ended question - approximately one minute.

 

    1. Place sensitive demographic questions at the end. This is especially true for financial questions such as household income. Mandatory requirements for answering sensitive questions will drastically increase the survey drop-out rate. If not necessary to the quality of the survey, include a "refuse" option.

 

    1. Use screeners wisely. Increasing the number of screeners can have an exponential effect on the cost of your panel. With each additional screener, the overall incidence level of the survey drops. The overall incidence level is calculated by dividing the total number of completed surveys by the total number of people who attempted the first question of the survey. Thus, if it took 500 people to fully answer 100 questionnaires, the survey incidence level is 20%.

 

    1. Limit the number of attribute ratings. Online survey respondents are in a hurry! Too many attribute ratings cause respondent boredom, inattentiveness, and straight-lining (selecting the same rating for everything). No more than 7-8 attributes is a good rule. Make sure the attributes are clear and concise. Carefully review each attribute to avoid redundancy. Any ambiguity can bias the results.

 

    1. Avoid ambiguity. Respondents are people who are kind enough to agree to participate in your survey. You cannot expect them to bear the burden of deciphering a poorly designed survey. All questions should be written clearly without any room for ambiguity. Ambiguity can lead to false results and ultimately poor business decision making.

 

  1. Utilize graphics. Graphics can help get the message across as well as reduce boredom. However, keep in mind that respondents have computers with different speeds, screen sizes and upload times.


Research Inc. is an Atlanta, Georgia full service market research firm offering a wide variety of research services, including focus group moderation, telephone interviewing, online surveys, and website usability studies.

Related: market research, Atlanta focus group research, professional focus group moderators.