How to Detect Survey Fraud
About 10% of responses to surveys may be disingenuous, according to Measuring U, a quantitative research firm based in Denver, Colorado. Such disingenuous responses are the result of many factors, including fraud, and they have a tremendous impact on survey quality, making it difficult to extract true insights or take action on survey results.
Say you’re a large enterprise organization that requires your locations to survey their customers for feedback. How can you be sure your location managers aren’t gaming the system, filling out surveys themselves trying to skew the results in their favor? Perhaps customers receiving the survey have been promised an incentive and are simply speeding through the questions or entering one-word answers to get to the prize. Either way, your survey results won’t provide an accurate reflection of your service quality.
Here are several practices that can lead to inaccurate survey results:
- Speeding: Respondents may speed through, not giving much thought to their answers.
- Straight-lining: Remember answering C to all those multiple-choice questions on that test you didn’t study for? Survey responders do this, too. Responses that have all the same answers (straight-lining) or answers in a pattern (zigzag or Christmas-tree) are an indication that the responder didn’t provide genuine responses.
- Skipping questions: Ignoring every single question not marked with an asterisk might be indicative of a lack of interest in providing genuine feedback.
- One-word answers: Not only do one-word or extremely short answers show disinterest, they also provide no usable insight.
Fortunately, advanced solutions such as Reputation Surveys offer ways to filter out and remove suspect survey responses, enabling organizations to improve the quality of their data and extract meaningful insights from their survey analysis.
How Do You Determine Response Quality?
Reputation.com Surveys uses four dimensions to analyze the quality of survey results — Time, Length, Straight-lining and Skipped Questions. You can choose which surveys to tag as suspect using any combination of the following characteristics:
- Time spent on answering each question is below the expected threshold.
- Comments are shorter than the determined length (e.g. 25, 50, or 100 characters).
- Multi-choice questions from a single respondent have the same answer (straight-lining).
- All questions on a survey that don’t require an answer are skipped.
When a survey is flagged as suspect, Reputation.com Surveys displays the “Suspect” tag next to the survey result in question, allowing you to investigate further.
Suspect responses can be removed from your survey analysis and results, but if you review the result and determine it is valid, you can choose to manually unmark if from to “Genuine.”
Reputation.com Surveys also allows you to adjust settings and filters to meet your specific requirements. For example, if you expect a minimum of 100 characters for responses to an open-ended question, you can filter out shorter responses.
Know Your True NPS and Take Action with Accurate Results
By identifying suspect survey responses and enabling you to decide whether a response is genuine or fraudulent, Reputation.com Surveys enables you to pinpoint sources of fraudulent responses, eliminate ingenuine responses from your analysis and obtain a truer, more accurate understanding of customer sentiment about your business. With this level of insight, you can make a plan of action to address problem areas and standardize on what’s working well across your organization.
Reputation Surveys delivers actionable data and enables you to continually improve the customer experience. Learn more about our Surveys solution here.