Some frequently cited concerns include:

• “It just isn't as accurate”
• “Online respondents are different from my customers”
• “I never know if the sample is full of ‘professional respondents’”
• “I don’t believe online samples are really representative of the U.S. population”

But, the lure of conducting research online is obvious, including timeliness (in some instances, can be conducted in a fraction of the time of other methods) and cost (frequently a lower cost option). Of course, potential pitfalls exist when utilizing online panels for conducting research. In addition, clients and researchers alike have to be well versed with the techniques to gain the best results from online methodologies.

Pick the right panel: There are a variety of sources to go to for your online survey needs. If using an online panel, chose the right panel for the right objective. One should know how the online panelists are recruited, rewarded and managed. Ask many questions, and feel free to specifically request panels that you have used in the past that have yielded good results.

Make the questionnaire engaging: Researchers and clients alike have a responsibility to the respondents. Do not make the questionnaire too long. Resist, if possible, long (boring) lists. Take advantage of the power of the Internet to design an interactive survey. Engage the respondent. Simple steps like these can go a long way in gathering more accurate and insightful data.

“Validate” the respondent: While we believe that consumers from online panels are generally truthful, responsive, and accurate, it is always best practices to insert some questions in a questionnaire to make sure that respondents are “tuned-in”, alert and paying attention to the survey. From “red herring” questions, to low incidence brand checks to “fake brand” identification to “pattern analysis”, a number of techniques can and should be utilized to ensure that you have the most accurate data possible. After all, multi-million dollar decisions are often made based on some online research. It is only wise to make sure that all steps have been taken to assure data soundness and data quality.

The full Paper is available in PDF format

No part of these materials may be reproduced in any form without written permission.