HSM E-briefing
April 2006

10 Questions to Ask
When Conducting an Internet Survey
[continued]

4.  Will we get a representative sample?

Internet access and familiarity with Internet navigation is very common but not universal.  Seniors, low-income groups, non-English speakers, and other demographic segments have lower-than-average rates of Internet use.  Even for business-to-business surveys there can be internal corporate firewalls or other obstacles to access and response.  (At HSM we often provide an Interactive Voice Response (IVR) option for those who may not have Internet access during work hours or who just prefer the alternative.)

With that said, for most consumer and business segments, Internet access and familiarity are not barriers to getting a representative sample.  Still, a targeted survey that depends on invitations sent to a specific list of respondents (such as physicians, employers, or health plan members) will be much more likely to obtain a sample that reflects the appropriate population than will an open survey that has portals in open-access Web sites that can be completed by anyone.  For most research with consumers and business segments, a targeted survey among groups with demonstrated access to the Internet typically provides a very representative sample of responses.

5.  Can we handle the programming?

Some simple surveys do not require advanced features such as skip patterns, branches, or dynamic content (where text in the survey changes based on previous responses).   Programming a survey without such features can rely on generic survey tools that are available on a Web site or that can be downloaded to your computer.

However, advanced features can add greatly to a survey’s value and the ease of use by respondents.  The survey will look more professional and it can glean more useful information. 

To get these advanced features you will need advanced software.  There are several software packages available that claim to do complex Internet surveys “quickly and easily.”  But in our experience it is usually not quick and easy.  The learning curve is steep for these packages.  Even with a software program, you will probably need to know other languages such as HTML and JavaScript or PHP if you want dynamic content. 

You’ll also need to know the package well enough to handle errors when a respondent advances through the survey, and to spot errors in your code.  Also, a complex survey has a more complex data table, and the importance of interconnectivity between the survey and the database cannot be understated. 

If you plan to conduct frequent Internet surveys, it might make sense to invest the money and staff necessary to use a capable software package.  However, if you do not want to jump into the programming business, you will likely get better results in less time and with fewer headaches by using an established survey vendor for the programming.

6.  Is our survey database security up to par?

Do you know how to implement Secure Socket Layer (SSL)?  You might not need to know how.  Most software packages provide security features to avoid unauthorized access to survey administration or survey data.  However, you should still make certain that your server offers a high level of protection.  If you are renting space on a Web server owned by another company, you probably won’t need to worry about virus protection and firewalls.  But if you have your own dedicated Web server, or rent one from a vendor, you will need to be aware of and understand the basics about protecting the information on your server and transmissions from respondents.

Read #1-3 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