Data Science/Computational Social Science Seminar: Lia Bozarth

Date: 
Thu, 10/10/2019 - 12:00pm

Room 2245 North Quad

Profit for you and me: comparing ad servers on fake and traditional news sites

Abstract:

Existing studies demonstrate that many fake news sites have profited considerably through the combination of creating and publishing sensational false stories, attracting large viewer traffic and selling generous ad spaces. Consequently, motivated scholars and political pundits alike have urged popular ad serving companies such as Google’s DoubleClick to blacklist known fake news sites, aiming to undercut these publishers’ potential income. Yet, the effectiveness and the cost of this proposal is, thus far, understudied.

For this preliminary study, we first compare and contrast i) the distribution of the number of unique ad servers and ad URLs available on over 2000 fake and traditional news sites. We then ii) examine whether ad servers on fake news sites are less well-known and more maliciousness compared to those available on traditional news sites. Next, we iii) assess whether leading ad servers such as DoubleClick and Epom blocking fake news sites will indeed make a significant impact on these sites’ potential revenue. Finally, we iv) estimate the direct monetary cost for these ad servers if they choose to leave fake news sites.

Speaker Bio:   

Lia Bozarth is currently a 4th-year doctoral candidate at UMSI. Her research is focused on applying computational methods to study and address pitfalls and safeguards in online political communications. Her projects were published in both the academia and the press.