Attach3_SFTXT_Antifraud_11192012

SFTXT – Attachment 3

Attachment 3 - SFTXT Antifraud Measures

• Use of CAPTCHA. CAPTCHA is a technique used to verify that a person, and not a computer, is accessing the website. A common type of CAPTCHA asks a user to type letters or numbers that appear in a distorted image on the screen. This task is very difficult for a computer to perform, so it is a good way to verify that there is actually a person on the other end of the transaction.

• Honesty pledge. An honesty pledge can be implemented at the start of a survey to attempt to improve data quality. Users are basically asked to acknowledge that they intend to answer the survey truthfully. Results from a recent study presented at the American Association of Public Opinion Reporting suggested that these types of pledges can help reduce the occurrence of straightlining, question skipping, and other types of fraud.

• Ensuring that all IP addresses are U.S. based. Since many of the scams that have been identified in other recent research seem to have been coming from international IP addresses, restricting the sample to U.S.-based addresses may help to reduce fraud. (This is also a logical plan for this study since there may be complications with involving international cell service providers in the SFTXT program).

• Confirming enrollment via email. Confirming participants have a working e-mail address by sending them an e-mail and asking them to click a link within the e-mail will further confirm ability to participate in the evaluation and that the e-mail address is a valid one and not one someone made up to collect multiple incentives.

• Confirming working cell phone number. In addition to confirming that these are working phone numbers, Mobile Commons can also compare all phone numbers in the study to ensure there is no duplication, which would prevent one person from attempting to enroll twice. Mobile Commons will also notify RTI if any text messages bounce back.

• Hosting the survey on a secure site. The secure site means that only those with a password can gain access. This reduces the likelihood that a participant could tell his friends to take the survey to collect incentives. Participants need to enter the username and password that will be sent to them with the link to the follow-up surveys in order to access each survey.

• Conducting rigorous data cleaning. This will include looking for straight-lining, anyone who completes the survey in an extremely short amount of time, etc. More details can be found in the analysis plan section on data editing.

1