Googling 9/11: Following the history of Google results and assessing the impact of algorithm changes

Members

Antoine Brunel, Camila G. Dantas, René König, Carlo de Gaetano, Gabriele Colombo, Tommaso Venturini

Introduction

Google fundamentally shapes the way we experience the web by creating hierarchical link lists for specific queries. The more ambiguous a query is, the more significant is the information political impact of the search engine, since its algorithm has to reduce more information. The marketing field search engine optimization (SEO) has recognized this and aims at achieving high rankings for commercial websites. But what about political topics? How are they represented in search engines? Are there particular types of websites or world views that are favoured in the results? How dynamic are the rankings? Do algorithm changes significantly change established hierarchies? Govcom.org’s IssueDramaturg provides the possibility to get insights “into the continuing drama of changing engine placement” (Govcom 2007). It allows for saving and comparing Google results for specific queries over time. This has been done for the term “9/11” for the period 2007 until 2012. This keyword is well-suited to grasp the information political impact of Google because it is

● ambiguous: “9/11” can refer e.g. to the terrorist attacks 2001 (and multiple aspects of it), any other event dated on September 11, a calculation and many more potential contexts.

● significant: The probably most common usage of “9/11” is the association to the September 11 attacks 2001, an event which has significantly influenced global politics and culture. Also many Google searches seem to be related to this meaning of 9/11, since there is a significant increase of “9/11” searches at every anniversary of the incident.

● politically-loaded: The September 11 attacks did not only have massive political effects which are controversially discussed, also the event itself was described in very divergent ways. Most remarkably, the participatory opportunities of the Web has allowed to establish alternative accounts opposed to the “official version” represented by mainstream media, governments etc. The so-called “9/11 truth movement” has been trying to promote this alternative account of the event.

Given this background, it becomes clear that the first results for the “9/11” query are very contested. How is 9/11 “defined” here? Which account is represented more prominently and how do algorithm changes impact the ranking?

Research Questions

How do algorithm changes impact the rankings of websites?

Can certain categories (alternative vs. mainstream accounts) of websites benefit from algorithm changes?

Methodology

In this project we rely on a dataset collected by the DMI team with the IssueDramaturg. Google.com has been queried daily with the term “9/11” (in English and with the minimum possible personalization), collecting the first 97 results from May 2007 until May 2012.

This data has been analyzed with the following steps:

Data extraction

Identification of most central websites

Selection of websites for further analysis

Visualization of development

Analysis of the algorithm impact

Data and tools

● The data collected by the DMI team for the query 9/11 from 2007 to May 2012 with the IssueDramaturg.

● A list of the main updates of the Google indexing algorithm.

Preliminary Findings

Slides from intermediary presentation on Prezi.

Further Research

Literature