HSM E-briefing
April 2006

The HSM Group, Ltd. provides the following information in the hopes that it will help you, our valued clients and friends, in your efforts to conduct better research.

10 Questions to Ask
When Conducting an Internet Survey

The appeal and potential of Internet surveys for research is huge.

• Respondents send data in a useful format.
• There is little data entry or coding needed.
• The primary fielding cost is programming.
• Response rates are improved because respondents can complete the survey at any time from anywhere.

On the other hand, the Internet environment has quite a few aspects that can limit the success of the research.  Here are 10 important questions that should be asked when designing an Internet survey:
 

1.  Will a Web survey provide as much information as a phone survey?

The conventional wisdom is that it is easier for respondents to give detailed feedback by talking during a phone survey, rather than typing text into an open-ended query in an Internet survey. 

 In our experience, the detail that we get from Internet open-ends (as long as the questions are focused and well-worded) is complete and insightful much more often than not.  The negative aspect of typing is apparently offset by the anonymity and flexibility of the Internet environment, which also allows respondents time to carefully consider their answers.   

2.  Will a Web survey keep a respondent’s attention?

Keep it short, keep it direct, keep it focused.  We know the temptation to ask everything about a particular topic or product.  In the “point and click” Internet environment it’s easy to assume that another question will not add much to the survey length. 

An Internet survey that takes longer than 10 minutes is pushing the limits for many respondents, unless the subject matter is compelling or the survey has particularly interesting features.  Test your survey by having others who are not familiar at all with your company or product do the survey.  Time them.  Ask them if they became bored or wanted to stop, and why.  Within each screen, minimize the up-and-down scrolling needed to view more questions.  Never make users scroll to the right or left in a survey. 

You may have a survey that is just the right length, but will it keep them interested?  With Web pages, you have an unlimited number of options with colors, pictures, and even audio/video, but not so many as to slow the download time or to prompt the computer to reject it or send to quarantine.  Simply changing a font style or color and the page background can make a survey look more professional. 

3.  Will our Internet survey work for everyone?

You’ll need to test your survey with multiple platforms (XP, Windows 98, etc.) and with multiple Web browsers (Internet Explorer, Netscape, etc.).  Alternative browsers such as Firefox and Opera are becoming more popular and may be prevalent among your respondents.  You’ll want to make sure your survey appears the same and works the same within all environments. 

Another important aspect is the download time for each screen.  The server where your survey resides should be fast, of course.  Your screens should not be so weighed down by pictures or audio that loading time for respondents is cumbersome.  Your dropout rate will spike if respondents have to wait 15 or 20 seconds for each screen to load.  Build your survey with conservative assumptions about the connections that your respondents have and the speed of their computers.

Read #4-6 in this top 10 list
Read #7-10 in this top 10 list

---

To discuss how HSM can help your organization better understand what your customers want, call Jim Hendrix, Vice President of Research and Economics, at 800-776-8078, ext. 310

Other e-briefings