Using Google Insights to Assess Egypt’s Jasmine Revolution
Mar 01, 2011
After the fall of Egyptian President Hosni Mubarak, numerous commentators identified the prevalence of the Arabic chant “al-sha`b yurid isqat al-nizam” (the people demand that the regime be overthrown) in protests ranging from Morocco to Bahrain. Scholars and journalists are on firm ground when they assert that the prevalence of this chant across the region is indicative of, in the words of Rashid Khalidi, “eminently reasonable demands for freedom, dignity, social justice, accountability, the rule of law, and democracy” across the Middle East and North Africa. Opinion is far more divided, however, on what the future holds for countries such as Egypt and Tunisia where the old regime has fallen.
Identifying likely outcomes in fast-paced and dynamic situations like the unrest currently gripping the Middle East is always difficult. “Leaderless” revolutions, such as those in Egypt and Tunisia, are often particularly problematic because conventional tools of intelligence are of limited use: a satellite can estimate a crowd size, but it does not help to identify the ideas that will inspire and sustain protestors. Even high placed human intelligence sources may lack certainty as to who the political actors are that matter among the masses of demonstrators.
This article explores the use of a powerful tool of open source data analysis, Google Insights for Search, which offers unique advantages for gaining insight into these mass movements, and the ideas, thought leaders, and personalities driving revolution. The article applies Google Insights to the recent revolution in Egypt, showing how the tool allows analysts to gain intellectual purchase on three different facets of this “Jasmine” revolution, namely: the role of religion in post-revolutionary Egypt, the Muslim Brotherhood, and an electronic straw poll of likely presidential candidates.
Why Google Insights?
One of the central challenges to assessing the likely outcome of a mass social movement as witnessed in Egypt is that researchers lack tools for understanding the impact of the ideas and thought leaders sustaining the social unrest. This problem is exacerbated by the fast pace of events and their unprecedented nature (the very facets that make them so important) because both factors limit the degree to which observers can rely on history to predict what the future holds. Furthermore, because these are mass movements, journalistic anecdotes or intelligence from individual sources may be misleading or a poor tool for intuiting what the masses want, and how much they want it. In this particularly challenging research environment, tools that quickly survey a large cross section of the Egyptian population should be of enormous interest to social scientists.
Google Insights is a free service and allows researchers to conduct near-instant analysis of the search terms typed into Google’s search engine. While the search results of this new tool need to be more comprehensively tested and the robustness of its findings are open to discussion, an assessment of 30 days of searches from January 25, 2011 until February 22, 2011 provides some counterintuitive conclusions that are explored in more detail below.
Traditional Measures of Public Attentiveness
Why are internet searches relevant? What additional insight does this tool provide researchers to developments on the ground in Egypt? To answer these questions, it is essential to understand what Google Insights can and cannot measure and compare its strengths and weaknesses to conventional tools for taking the social temperature in Egypt.
It is important to note that Google Insights is different from a public opinion poll. While public opinion polls measure aggregate attitudes, Google Insights helps better understand public “attention.” This is defined by Jens Newig in his study of environmental regulation as “the scarce resources—time and others—that citizens willingly dedicate towards thinking about publicly debated issues.” This concept is frequently measured as a “relative intensity (resource employment per unit time) or as a ratio (resource employment dedicated to one issue as compared to another issue competing for attention).” Crucially, because Google Insights measures change over time and has a lag of only 48 hours, it provides a constantly updating view of developments on the ground.
To estimate public attention, researchers traditionally utilize two tools: public opinion polls and media analysis. Public opinion polls ask respondents about their “most important problem,” with responses aggregated, normalized and displayed against a number of pre-determined issues. This approach faces a host of problems exacerbated by the social upheaval sweeping the region, including high cost and difficulty of repeated polling. In an alternative method, media coverage-based approach, researchers code news stories as a proxy for public attentiveness and standardize two issues against one another. The problem with this approach is that it is difficult to estimate both the impact of the media and the direction of causality between public attention and media.
Measuring Public Attentiveness
Google, the dominant global search engine, provides real-time insight into individual “attentiveness,” but on a massive scale. This “database of intentions” has recently provided new opportunities for research across a range of different fields. Most notably, search data has helped scientists predict outbreaks of influenza, economists predict changes in car, home sales and unemployment levels, and sociologists predict consumer behavior such as attending movies or purchasing movies or video games. Search results are particularly appealing during moments of great upheaval (as long as the internet stays operational) because they can suggest results in near real-time at minimal cost compared to existing survey methodologies.
Google Insights allows users to compare interest in up to five terms over time. Results, sorted by time frame and location, are normalized, scaled and returned on a graph, which can be downloaded in a comma separated values (CSV) file for further analysis. Testing for convergent validity using vector auto-regression and correlation analysis, Ripberger compares Google Insights with a New York Times-based media attention measure and finds strong positive and statistically significant correlation coefficients for a variety of terms such as global warming, health care and terrorism.
Challenges to Validity
There are at least four significant challenges to using Google Insights to measure and identify the forces animating Egyptian internet searches and perhaps more broadly Egyptian society.
First, using internet search leads to a selection bias. While internet access is not universal in Egypt (and indeed there is great variation in internet access throughout the Middle East), this may not be a challenge in the long run if internet access continues its impressive growth. Furthermore, it is not necessary to wait for near universal accessibility if the population of interest is young people, who were widely credited with being the driving force behind the revolution and who also have higher levels of connectivity. While there is little reason to believe that certain segments of Egyptian society prefer Google’s search engine to other options, there is some risk of an echo chamber effect if there is no diversity (particularly political) in the demographic of Google search users.
Second, Google Insights only allows normalized data, so it is not clear how large the raw numbers of each search term are. While this can be mitigated somewhat by experimenting with different cross category searches, the findings may have little validity if the overall number of search terms is quite small. This is a major concern, but one that can be mitigated by additional experimentation and drawing on additional sources of data, particularly site traffic.
Third, the information about the intention behind the search—such as whether a person is searching for a political party for a class research project, out of curiosity or because of an interest in joining the party—is not known. It is important to stress the distinction between attentiveness and opinion. In their paper on consumer behavior, Goel et al. found a “wide variability in the predictive power of search…and substantial differences in the relative value of search data compared to alternative sources.” This is a key point to consider and should serve as caution for other researchers never to rely on Google Insights alone in their own research.
Fourth, in some countries individuals may be cautious about openly searching for texts or videos of opposition figures, particularly in countries where there is a high degree of government internet monitoring. Tunisia (until recently) stood out among Arab countries as an example. This is much less of a concern in the period of this study because of the collapse of internal security services in Egypt and because most of the searches examined were not in any way illegal. There is little reason to think, therefore, that Egyptians were deterred in what they searched for in the 30-day period of this study.
While all of these challenges show the limitations of using Google Insights exclusively to draw conclusions about Egypt’s future, with a firm understanding of the distinction between attentiveness and public opinion the authors feel that using this tool has identified some major insights into the ideas, thought leaders, and issues animating Egypt during this crucial period.
One major concern that is essential for understanding the implications of Egypt’s revolution is trying to determine the role religion will play in a post-Mubarak Egypt. This is a complicated question to answer because prominent religious leaders (particularly the shaykh of al-Azhar University, the epicenter of Sunni religious education) have long been viewed as tied to the political establishment. This meant that the most important independent Egyptian religious figures (until the revolution) resided outside of the country. Two prominent religious figures have returned home since the uprising: Amr Khaled, a young, charismatic televangelist, and Shaykh Yusuf al-Qaradawi, a television personality with his own show on al-Jazira and a popular and far more conservative religious authority. Perhaps not surprisingly given the preponderance of younger Egyptians on the net, a comparison of searches for these two religious figures shows unequivocally Khaled’s predominance online.
To determine whether this method is valid there must be some evidence that Google Insights is actually measuring ground truth in Egypt—in other words, that daily events in Egypt are impacting overall search levels. A number of tests confirm that Google searches are reflecting interest in these two popular religious figures in Egypt. For example, a sharp spike in searches for the prominent conservative cleric Yusuf al-Qaradawi on February 18 is undoubtedly a reflection of his first public sermon in Tahrir Square, watched by millions of Egyptians and largely interpreted by commentators as a sign of his prominence and even preeminent position among religious scholars in Egypt. While al-Qaradawi is undoubtedly an influential figure, examination of search data from January 25 to February 22 demonstrates that his influence is not particularly significant on the internet and searches for Islamonline (in Arabic script and English), a site associated with al-Qaradawi, were also flat.
Instead, it is the young, charismatic preacher Amr Khaled who currently dominates the Egyptian internet. Confirming research by the Washington Post on Facebook activity after the revolution, Khaled is undoubtedly the most popular religious figure on the internet in Egypt. In the roughly 30 days of this study, Khaled had five times more searches for his name in Arabic script than al-Qaradawi. Such was the surge of attentiveness to Khaled that he actually surpassed searches for the Lebanese pop star Nancy Ajram—a result that suggests the genuine and intense amount of attention Khaled’s charismatic and youth-oriented religious message is generating in the post-revolutionary period.
The Muslim Brotherhood
As demonstrations swelled across the country, numerous articles speculated about what role the Muslim Brotherhood would play in Egypt’s future. Particularly given the Brotherhood’s lengthy experience in opposition, its legions of supporters and established leadership, most authors presumed that the Muslim Brotherhood would be the most significant and influential party in the Egyptian opposition. Because of the Brotherhood’s history of violence in the 1950s and 1960s and emphasis that Egypt should be governed by Shari`a (Islamic law), some authors viewed an Egypt dominated by the Brotherhood as diametrically opposed to the interests of the United States as well as Israel.
While the Muslim Brotherhood is an experienced and important force in Egypt’s political opposition, there is little evidence to suggest that the group’s ideas and agenda are attracting significant attention online in Egypt. This finding should be tempered by the knowledge that the Brotherhood has long had access to newspapers and television, and therefore the internet is only part of how the group spreads its message. A further factor may be that unlike other opposition groups, the Brotherhood, because of its identifiable hierarchy and existing infrastructure, actually has resources to lose if it makes a strategic error, and therefore is taking a cautious approach until political upheaval subsides. In the words of one long-time observer of the Muslim Brotherhood, “The group has very long time horizons and remains committed to a long-term goal of bottom up social change.” In an era of suddenly wide-open political possibilities, however, the Brotherhood’s innate conservatism may be working against it (at least online) as Egyptians look for new political leaders and parties.
The findings from this study cut against the characterization of the Brotherhood as a behemoth among a disorganized opposition and bolsters a far more cautionary tone on the group’s capacities and influence expressed by experts who have extensively studied the group. While the Muslim Brotherhood may be powerful on the ground, its virtual presence is not generating significant levels of attentiveness. Searches for prominent leaders of the Brotherhood—including spokesmen Essam el-Errian and Abdel Moneim Abou el-Fotouh, Supreme Guide Mohammed Badie, former Supreme Guide Muhammad Mahdi Akif, and other leaders including lawyer Sobhi Salih and former MP Mohammed al-Baltagi—did not return enough results for Google Insights to graph. Additional searches for “the brotherhood” in Arabic and “ikhwanonline” (the Muslim Brotherhood’s main website) showed only modest results after the Egyptian government restored the internet on February 1.
The recent news that a younger cohort of Brotherhood members, many of whom were the first to join the protests (against the wishes of more senior party members), are now openly discussing forming their own political party could provide an opportunity for future analysis about the impact of this generational rift online. The Brotherhood may also lose supporters to the breakaway Wasat Party, founded by former Brotherhood member Abu al-`Ila Madi, which received a notable spike in searches on February 19 when it was recognized as the first official religious party in post-revolutionary Egypt.
Finally, contrary to speculation that the Brotherhood would receive a boost by delegating former International Atomic Energy Agency (IAEA) head Mohammed ElBaradei to lead negotiations with the government and his robust defense of the right of the Muslim Brotherhood to contest elections, ElBaradei did not figure prominently in online searches of opposition figures. A comparison of searches for the leading opposition figures including Mohammed ElBaradei, Ayman Nour, and Amr Moussa shows Moussa fast outpacing other rivals. While ElBaradei generated more searches than opposition figure Ayman Nour, he was outclassed by the current head of the Arab League, Amr Moussa, by a ratio of 5:1. This is a real testament to the success of the strategy employed by the Mubarak government to get the charismatic Moussa out of the public eye by relegating him to the Arab League. Yet with the Mubarak regime gone, Moussa is clearly the opposition leader who people are most interested in and his search results lend support to reporting that he is currently a front-runner for the presidency.
Separately, the significant increase in attentiveness to well-known figures such as Amr Moussa was absolutely dwarfed by searches for two individuals who were virtually unknown less than a year ago: Khalid Said and Wael Ghonim. Searches for Khalid Said, a young man allegedly killed by police officers in Alexandria in 2010, and Wael Ghonim, a Google marketing executive who founded a Facebook page commemorating Khalid (We are All Khaled Said) and was detained for a period of 12 days by government security officials, were both “breakout” searches for the period of study. While there is evidence that searches for these two individuals may be trailing off, they were the most popular of the search terms tested and demonstrate two key points: Egyptians do turn to Google’s search engine to find out more about the personalities and figures that they hear about in the news, and the surge in searches demonstrates the internet’s power in Egypt as an increasingly important alternative source of information, even in a country saturated with satellite channels and a raucous free press.
Avenues for Future Investigation
As with all major shifts in foreign affairs, opinion among experts seems to be polarized into two main camps around the question of “whither Egypt?” In the first camp are the optimists: those who view the fall of the Mubarak regime as a Berlin Wall type event, nothing less than a sea change in the politics of the region. These authorities view Egypt’s peaceful revolution as part of a broader “Arab Spring” and note that political change has irreversibly arrived in the region. Even the most hardened dictators such as Libya’s Mu`ammar Qadhafi and Yemen’s Ali Abdullah Salih have been put on notice. These same authors note that Egypt’s Muslim Brotherhood and the conservative al-Nahda party in Tunisia have played only a minor role in the revolutions in those two countries and big ideas such as hurriyya (freedom) and dimuqratiyya (democracy) are the watchwords of the young and jubilant protestors in Tunis and Cairo, not Muslim Brotherhood slogans, much less the jihad called for by al-Qa`ida.
At the other end of the spectrum are those experts who tend to deemphasize the significance of the ancien regimes’ fall from power. These skeptics of the “Arab Spring” thesis note that in Egypt the military is still in charge, the government consists mainly of Mubarak appointees, and Egyptians are unlikely to remain enamored with democracy when faced with rising food prices and continuing political instability—only able to resort to weak, fractionalized opposition parties with little power to address Egypt’s pressing political problems. This more cautious camp tends to accept as fact that the Muslim Brotherhood is the best organized and most popular opposition movement, that Egyptian military officers enjoy economic and political benefits that they are unlikely to give up easily, and that jihadists such as Ayman al-Zawahiri and Atiyah Abdul al-Rahman are undoubtedly recalculating their communications strategy to take maximum advantage of the political opportunity provided for them by this new and uncertain future.
Numerous scholars have written about how hard it is to predict the type of unprecedented events like the Tunisian and Egyptian revolutions. Those who have studied this issue from the perspective of the intelligence community (who are often the first to be excoriated for failing to “predict” these events) note that events of this magnitude are often missed because analysts rely far too much on history as the guide to what the future holds. Particularly when looking at events that are literally unprecedented, history’s predictive power is considerably reduced.
A second issue of concern is the so-called “signal to noise.” This is the problem of identifying insightful and relevant data points (signal) from a sea of information (noise). This process typically occurs only after the revolution, war, or terrorist attack has taken place, with the U.S. Congress or the media holding intelligence officials’ collective feet to the fire for not predicting with enough accuracy that the Japanese would attack Pearl Harbor or that the shah would fall in Iran. The problem is that information viewed retroactively always points far more definitively to a decisive outcome than in the days before the incident actually occurs.
This longstanding debate on “intelligence failure” does not even touch the fraught problem of what to do about the immediate future. While critics are pressing the intelligence community about how they “missed” Egypt, these same experts are undoubtedly struggling to provide relevant insights into how developments will unfold now that the protestors have ousted the dictator and the cloak and dagger politics of the post-revolution are in full swing. Will religious fundamentalism grow in Egypt? Will the Muslim Brotherhood dominate a fractured and ineffectual opposition? Now that the hated security apparatus is on the defensive, will jihadists have the opportunity to regroup?
These are daunting, but crucial questions. Research using preliminary experimentation with Google’s Insight for Search service supplemented with extensive discussions with regional experts and methodological testing has provided some interesting and counterintuitive results about the ideas and personalities currently shaping Egypt’s future. While there are still numerous methodological challenges to this approach, Google Insights presents real promise as a new tool for overcoming some old and enduring challenges as researchers and academics try to address the all important question: “whither Egypt?”
Joshua Goldstein is a technology consultant and Ph.D. candidate at Princeton’s Woodrow Wilson School studying the links between telecommunications, non-state threats and democratization. He has worked extensively in East Africa, most recently leading the State Department’s Apps4Africa competition. While completing his master’s degree at The Fletcher School, he conducted research with Google’s Public Policy team and Harvard’s Berkman Center for Internet & Society. His work has appeared in Princeton’s Journal of Public and International Affairs, The Fletcher Forum of World Affairs, Harvard’s Berkman Center for Internet & Society Working Paper Series and other publications.
Gabriel Koehler-Derrick is an Associate at the Combating Terrorism Center and an instructor at the U.S. Military Academy, West Point. Mr. Koehler-Derrick holds an M.A. in International Affairs from Columbia University’s School of International and Public Affairs and a B.A. in International Relations from Tufts University.
 The authors would like to thank Nelly Lahoud, Arie Perliger and Steven Brooke who all provided substantive feedback and comments.
 Rashid Khalidi, “Reflections on the Revolutions in Tunisia and Egypt,” Foreign Policy, February 24, 2011.
 Yahoo Clues currently only offers data on U.S. searches. Bing’s “social” is a comparable tool but follows social media sites Twitter and Facebook, not overall search trends. Bing also does not allow users to filter results by country.
 All analysis was conducted exclusively on searches in Egypt, using Arabic script except where noted. According to Alexa (a website that measures traffic), google.com.eg is the second most popular website in Egypt. Google.com is the fourth. It is important to note that the Egyptian government severed internet access in the country from January 25 to February 1-2. During that time period, no data could be collected.
 Jens Newig, “Public Attention, Political Action: The Example of Environmental Regulation,” Rationality and Society 16:2 (2004).
 Joseph Ripberger, “Capturing Curiosity: Using Internet Search Trends to Measure Public Awareness,” Policy Studies Journal (forthcoming).
 Stuart N. Soroka, “Issue Attributes and Agenda-Setting by Media, the Public, and Policymakers in Canada,” International Journal of Public Opinion Research 14:3 (2002).
 Jeremy Ginsberg et al., “Detecting Influenza Epidemics Using Search Engine Query Data,” Nature 457 (2009).
 Nikolaos Askitas and Klaus Zimmermann, “Google Econometrics and Unemployment Forecasting,” Applied Economics Quarterly 55:2 (2009).
 Sharad Goel et al., “Predicting Consumer Behavior with Web Search,” Proceedings of the National Academy of Sciences in the United States of America 107:41 (2010).
 Internet access and use has dramatically increased during the past five years in Egypt. For statistics, see http://data.un.org/Data.aspx?q=egypt&d=ITU&f=ind1Code%3aI99H%3bcountryCode%3aEGY.
 A 2008 study by the Egyptian government provides some basic data on typical internet users in Egypt. For details, see “The Future of Internet Economy in Egypt,” Egyptian Ministry of Communications and Information Technology, 2008.
 Searches for Shaykh Tantawi, the long-standing leader of al-Azhar, show no significant activity except for a spike in searches in the month of his death in March 2010. There were limited searches for the grand imam of al-Azhar, Ahmed al-Tayib, in the 30-day period of this study.
 David D. Kirkpatrick, “After Long Exile, Sunni Cleric Takes Role in Egypt,” New York Times, February 16, 2011.
 “A Social Foothold in Egypt,” Washington Post, February 3, 2011.
 Amena Bakr, “Major Muslim TV Preacher Amr Khaled Heads for Cairo,” Reuters, January 28, 2011. In Khaled’s case, his search results were almost certainly boosted by his return to Egypt on January 28 (in the midst of the internet blackout) and his revelation that he was kept from publicly preaching in Egypt by the Mubarak regime.
 Nathan J. Brown, “Will Slow and Steady Win the Race?” Foreign Policy, February 28, 2011.
 David D. Kirkpatrick and J. David Goodman, “Muslim Brotherhood in Egypt to be Political Party,” New York Times, February 15, 2011; Richard Allen Greene, “Egypt’s Muslim Brotherhood: A Force to be Feared?” CNN, January 31, 2011.
 Sallai Meridor, “What Israel Fears in Egypt,” Washington Post, February 9, 2011.
 Personal interview, Steven Brooke, February 20, 2008.
 The results of the overwhelming approval of the amendments to the constitution in late March might call into question this assessment. Numerous commentators credited the Brotherhood with the approval because of the significant push urging its followers to vote “yes” on the amendments. That being said, the “yes/no” nature of the referendum and the variety of competing rationales for voting either way make it difficult to interpret the Brotherhood’s overall impact. Searches for “Constitutional Amendments” for the period of March 16 to March 20 were well below searches for “Japan” and “Libya” (in Arabic script).
 Lorenzo Vidino, “Five Myths About the Muslim Brotherhood,” Washington Post, March 4, 2011; Steven Brooke and Shadi Hamid, “The Muslim Brotherhood’s Role in the Egyptian Revolution,” CTC Sentinel 4:2 (2011).
 Because of the weakness of political parties under the Mubarak government, searches for political parties (with the exception of the Brotherhood) turn up far fewer results than prominent individuals.
 Issandr El Amrani, “Egypt and Tunisia’s Unfinished Revolutions,” Time Magazine, March 6, 2011.
 Nobel laureate Ahmed Zewail and Tariq Bishri, the historian and legal scholar in charge of the Constitutional Reform Committee, were also included in this test but returned few results.
 This supports polling results from a Pechter Middle East phone poll cited by The Washington Institute for Near East Policy taken from February 5-8, 2011.
 Amr Moussa’s prominent role in the Arab League leading the discussion of a “No Fly Zone” in Libya means that this gap in attentiveness has only increased in late March.
 According to Google, this indicates growth greater than 5,000%.
 Roger Cohen, “From Oklahoma to Tobruk,” New York Times, February 24, 2011.
 George Friedman, “Egypt: The Distance between Enthusiasm and Reality,” Stratfor, February 14, 2011.
 Michael Scheuer, “Why the Middle East Revolts Will Help al-Qaeda,” Washington Post, March 4, 2011.
 “Why Didn’t the U.S. Foresee the Arab Revolts?” New York Times, February 24, 2011.