In our last article, we defined survey fraud and provided you with ways to spot it. Identifying survey fraud is the first step you need to take to prevent it.
But knowing how to spot survey fraud is only half of the process. For the second half of our survey fraud discussion, you must learn different ways to prevent it. In this article, you'll discover how you can transform that understanding into accurate survey data.
1. Check consistent responses
Whether bots or scammers take surveys, question answers are often inconsistent. This is because speed is often prioritized over accuracy to get through more surveys (leading to more rewards). By including the same question two or three times in a survey, you can check for consistency. If answers to these same questions differ, you can remove that survey from your data pool. Having multiple of the same questions is one example of "effort checks," as they see how well the person is paying attention.
2. Cross-reference any identifying information
Identifying data lets you avoid repeat surveys. Typically, this identifying information is limited to an email. However, some survey information (like census data) also asks for names. When cross-referencing data, you generally want to find fake data (someone making up an identity) or duplicate data (someone using the same credentials across multiple submissions).
Other forms of identity information include IP addresses. If the scammer is coming from the same IP address multiple times, you can prevent this. However, thanks to modern software, you don't want to limit your identity checks to IPs. Always include additional data of some kind.
3. Use personalized URLs
In cases where your survey asks anonymous questions, you can rely on personalized URLs. This way, it keeps the survey data you gather anonymous, but allows you to prevent duplicate surveys. When someone uses the duplicate URL once, this URL is shut down. In this way, you prevent multiple submissions from the same person.
4. Ask pre-screening questions
Pre-screening questions provide two benefits.
First, they let you narrow your data groups down to specific demographics. Doing this limits your data gathering to those groups, making your data more precise.
The second use of pre-screening is the ability to limit takers from going through the entire survey if they are a bot or someone wanting rewards. Combining this with other techniques (asking multiple questions, having personalized URLs, etc.) creates a larger time requirement, which makes your survey less lucrative for scammers.
5. Use a CAPTCHA
To prevent automated software from taking your survey, the ultimate method for doing so is still using a CAPTCHA. Because bots still struggle with reading visual or audio data, this prevents automated systems from using your surveys for rewards. A CAPTCHA won't prevent humans who participate in survey fraud. So, using this tool by itself is an ineffective tool. Always couple it with human-prevention tools, like effort checks, to make your fraud checkers comprehensive.
6. Time the survey
Real respondents will take time to consider and react to your questions. However, if you time the survey and see that the participant has rushed through, that's a sign of a fraudulent survey respondent. Take the time to test how long it should take to complete your survey using a quick testing phase. If survey respondents complete the survey within an acceptable threshold (one minute, 30 seconds, etc.), the survey is valid. But if it takes them much shorter than average, that's a sign that you might be dealing with a fraudster.
7. Ask a few open-ended questions
Open-ended questions are ideal for identifying bots and acting as an effort test. Asking for survey comments is one example, as it's a sign that the respondent is considering your questions. If you are trying to focus on gathering numbers-based data (quantitative), these questions might not be part of your initial gathering efforts. Again, these effort questions are not meant to collect information but to identify a valid respondent.
8. Use device fingerprinting
Online advertisers use device fingerprinting as a way to identify the user. This, in combination with other data, is a way to overcome software built to hide your IP.Device fingerprinting in combination with the zip code and email address can help identify potential fraudsters.
9. Use voice-based data collection systems
In our previous article, we identified several reasons why voice-based systems are a great way to help identify survey fraud. This is because vocal responses are much harder to fake.Vocal responses also often require more thought, making it more viable to gather and sort data from open-ended questions. By leveraging the latest technology, you can outpace the pitfalls that come from survey fraud.
True Reply is a proven system for gathering voice-based qualitative and quantitative data. Leveraging our systems will allow you to prevent survey fraud in your studies and get more robust responses. The evidence of its use can be found across 250 studies and over 300,000 responses collected through this platform.
To find out how True Reply can help your data become truly accurate, contact our team today.