How to guarantee the
quality of online surveys

By Maxime Le Bescond,
VP Sales, Dynata France

While most survey respondents are engaged and honest in their answers, those who are not should be flagged and discarded.

In the context of a survey or poll, technology enables us to implement increasingly thorough checks. Verification procedures during telephone or face-to-face interviews – both at the time of the interview and at the time of data entry – are usually conducted to ensure the legitimacy of the interviewer, rather than that of the participant.

Without these verification procedures in place, there would be numerous cases of fraud, with ill-intentioned individuals posing as survey institutes in order to obtain sensitive information from respondents. In the case of online surveys, the risk is reversed: participants have the opportunity to lie, particularly about their identity, in order to take part in an incentivized survey.

How do you ensure the identity of your panelists?

Ensuring the identity of your panelists is crucial for any data provider or survey institute, as the accuracy of your results depends on it. However, when it comes to incentivized polls or surveys, attempts at fraud are rife, especially when the panel selection takes place in a part of the world where small monetary incentives represent a significant economic value. For this reason, a growing number of online survey institutes are beginning to restrict the perimeter of IP addresses accepted as part of a survey.

Technology must also be able to prevent the same person from taking part in the same survey more than once, to ensure that the results are not altered. To discourage participants from cheating when taking part in a survey, it is recommended to indicate the topic of the study and its remuneration only after the initial selection process, in order to minimize the risk of fraudulent qualification.

In addition, recruiters can subsequently act in concert with technological solutions to prevent fraudulent responses from slipping into questionnaires and pre-questionnaires. For example:

  • By recruiting participants through trusted partnerships that include loyalty programs whose members are controlled at source
  • By using retroactive solutions to erase any suspicious responses reported by a client
  • By imposing the two-factor authentication required to receive rewards
  • Via long-term engagement with panelists.

Traditional methods are sometimes the most appropriate, such as asking a panelist for information they are supposed to know (such as their age or gender), and checking the consistency of their answers to assess whether they are being honest on other points. The questionnaire itself can reveal whether the respondent is indeed who they say they are, by asking for information they should know based on who they say they are. This is easier to implement in a B2B survey, where questions based on professional knowledge can be asked.

Nevertheless, we shouldn’t “see fraud everywhere”. It’s very easy for a genuine participant to click incorrectly in an online survey, and mistakes are far more common than malice, unless there is ample evidence to the contrary.

For panels, establishing a relationship of trust and commitment, together with new and more effective levels of verification, is the best guarantee of quality results in online surveys.

This can be achieved through the design of open-ended questions and the analysis of responses, as it is usually difficulty to make up complex, unique and relevant answers to open-ended questions that require specific knowledge and significant engagement.

You might also be interested in:
A shared vision of data quality
Infographic: The Dynamic Approach to Data Quality with Dynata

Any questions for our experts? Get in touch here >


Article published in the Journal du Net on 26 September 2023:
to read the original version in French, click here >