When everything is set-up a thorough testing process should take place, a full dress rehearsal before the survey finally goes live. Here are the steps we recommend:

• Technical testing of the online survey. As a minimum your survey should operate identically in MS Internet Explorer versions 5 and over as well as Mozilla based browsers such as Firefox and Safari on the Mac. This will give you around 95% coverage of the browsers currently in use. To check on the latest browser usage statistics visit:

  http://www.w3schools.com/browsers/browsers_stats.asp

• Technical testing of the email communications. We recommend validating your emails in Outlook, Outlook Express, Gmail and Hotmail. If you are running a business to business study then you should also test your email in Lotus Notes as this is used by many large organisations and often has trouble displaying HTML emails as intended.
  
  
• Pilot group. If you are running a staff or customer survey or similar then creating a small pilot group of a handful of people and running the study just for them is a good way to test not only the technical aspects but also the content of the study. If any glaring errors or omissions have been made in the questionnaire then they may well come to light here giving you a chance to deal with them before going live.
  
  
• Data testing. When the pilot has been run the data generated should be reviewed for anything that might appear out of the ordinary. Is there evidence that certain questions are being skipped? if so the routing will need to be checked as set-up errors here can sometimes slip though unnoticed. Its also worth having a look at the verbatims just to see if any comments have been made about questions not making sense or options having been missed from lists and so on.

When these steps have been completed then any test data should be cleaned out.

Full testing as described above is provided as part of the ORS Standard Survey Package.