Information Safety And Accuracy: Is Google Trustable?
This paper considers information safety and accuracy in the digital age using Google as an entry point. In doing so, it explores the role media play in shaping the relationship of information, privacy, and trust between Google and the public.
Photo credit: Tyler Olson
This inquiry is undertaken using framing theory to guide a content analysis of the way Google is presented in New York Times articles from a two-year period ending in November, 2005.
Analysis of the extensive coverage of Google's share price and earnings reports leads to the conclusion that trust in Google is fostered in part simply by reports of its fiscal success. To the extent that this is true, meaningful public debate about information policies is inhibited.
by Lee Shaker
On any given day, more than 60 million American adults use an online search engine -- a number that is rapidly increasing. Among search engines, Google is visited the most -- by as many as 89.8 million unique visitors in one month.
Google's rising traffic, continuing development of successful business models, and improving accuracy and depth has significant implications for the way that the Internet is experienced and organized. Beyond this, as Google grows entrenched as a central hub of Internet use, the company's role as an arbiter of information becomes increasingly important.
In the shift to a data environment dominated by the Internet, there are important questions about information integrity -- the credibility of information provided and the safety or privacy of information collected. No matter what the source, it is important for the audience to know the provenance and accuracy of information that is provided.
In two-way exchanges that occur on the Internet, it is also necessary for data collected about individuals to be used appropriately. In evaluating information from traditional media outlets, certain expectations and understandings about reporting, objectivity, and bias have been established over time.
Because of this, a shorthand assessment of credibility may involve simply trust in the institutional framework of the media as a whole, the individual outlet at hand, or even the author of a single story in the place of repeated critical inquires. In other words, capital developed over time by the media source formed the reputational trust that legitimized information.
This capital is still in its embryonic stages in the rapidly evolving online mediascape. The structure of a media institution in the online world is still in flux, but, for online media companies, it is vitally important to cultivate stable relationships with users.
For example, in its free-to-the-public form, Google is totally dependent on providing a critical mass of users to advertisers. Since Google traffics in public information, the audience must believe that the company serves up 'better' information than competitors and protects users' interests. Put simply, to draw the audience necessary to flourish, Google must cultivate the public's trust.
[in original, this is a highlighted quote] On any given day, more than 60 million American adults use an online search engine ...
An extensive body of scholarly literature relates to privacy, trust, and new information media. This work rises from an array of perspectives, among which are: cultural theorists (Elmer, 2004), communication scholars (Danna and Gandy, 2002; Phillips, 2003), sociologists (Lyon, 2003), computer scientists (Riegelsberger, et al., 2003), and marketers and economists (Kracher, et al., 2005; Cheskin Research, 1999).
Drawing upon these works, this paper seeks to explore the role media play in shaping the relationship of information, privacy, and trust between Google and the public. This inquiry is undertaken using framing theory (Pan and Kosicki, 1993; Scheufele, 1999) to guide a content analysis of the way Google is presented in New York Times articles from a two-year period ending in November, 2005.
While media coverage is far from the only component in the formation of trust, it does contribute to the process. Analysis of the extensive coverage of Google's share price and earnings reports leads to the conclusion that trust in Google is to some extent fostered simply by its fiscal success.
Information integrity and Google
Though most of its services are nominally free to the consumer, Google is an e-commerce vendor. The advertiser-driven model that the company adheres to mimics that of broadcast radio and television stations: create and convey information product to an audience, capture their attention, and package their eyeballs for sale to advertisers.
Though similar to broadcast businesses, Google clearly diverges from this model in one critical way: the nature of the Internet requires information to flow two ways, placing Google in a position to collect vast databases of information describing who and how people use their services.
The virtual nature of interactions between individuals and Google causes the trust relationship to be less intuitive than those developed in the physical world -- no physical goods change hands, no handshakes to base faith on.
Trust in this disembodied exchange operates in two areas of information integrity. First, the service rendered must satisfy the promise of a service offered. Put in other words, a search must return accurate, credible information quickly.
Second, the ancillary details of the transaction must not be harnessed by Google in a manner that damages the user. Because each interaction spawns an array of transaction-generated information (TGI), users rely on Google to act appropriately after the fact. As John Battelle puts it,
"That's a pretty large helping of trust we're asking [Google] to ladle onto their corporate plate. And I'm not sure either we or they are entirely sure what to do with the implications of such a transfer. Just thinking about these implications makes a reasonable person's head hurt." (1. source: p.15)
In the following two sections, research describing privacy concerns and the formation and importance of trust is explored to provide context for my subsequent analysis.
Photo credit: Michael Osterrieder
"Complex technology that involves the collection, processing, and sharing of information about individuals and groups that is generated through their daily lives as citizens, employees, and consumers and is used to coordinate and control their access to the goods and services that define life in the modern capitalist economy." (2. source: O.H. Gandy, The Panoptic Sort: A political economy of personal information. p.15)
In a word, this "dataveillance" process is embodied by the Internet. Here, data collection, processing, and sharing merges seamlessly with the daily actions of citizens, employees, and consumers as interactions occur between the server and terminal computers.
Each user has the capability to methodically record precise records of her actions; at the same time, each vendor has the ability to collect data on every interaction.
Each user has the capability to methodically record precise records of her actions; at the same time, each vendor has the ability to collect data on every interaction.
Though both user and vendor have the ability to record TGI, the usefulness of this ability is not equal for individuals and institutions. It may be convenient for an individual to automatically record the articles she reads or archive the purchases she makes -- possibilities made feasible by Google Desktop and Gmail.
A particularly astute computer user may even want to monitor how much time he spends with certain applications or Web sites -- but so far this requires specialized software that is not widely available.
Recording data is not difficult -- to a certain extent, it happens by default -- but to make sense of this data, sophisticated analysis is necessary. It is this step that reveals the differences in the utility of databases for individuals versus institutions.
While the individual is largely restricted to manually culling through the limited data that describes their behavior, institutions have terabytes of data stored and vast amounts of processing power available to churn through it.
This scrutiny results in social sorting that, "Sieves and sorts for the purpose of assessment, of judgment. It thus affects people's lifestyle choices -- if you won't accept the cookie that reports your surfing habits to the parent company, don't expect that information or access will be available ... ." (3. source: D. Lyon (editor) Surveillance as Social Sorting: Privacy, risk, and digital discrimination
The utility of data describing the audience is twofold for information vendors like Google. First, greater knowledge of each consumer can precipitate price discrimination.
The ability to effectively price-discriminate promises an increase in profit. By ascertaining what each potential consumer is willing and able to pay for a product and adjusting the price accordingly, vendors gain the ability to wring each drop of profit out of every interaction.
The second benefit relates to the other available revenue stream for media companies: advertising. As a commodity, "The audience has scarcely any value if it cannot be defined in some way (in terms of demographic make-up or at least in terms of size)." (4. source: Arvidsson, "On the 'pre-history of the panoptic sort: Mobility in market research")
Dataveillance yields a wealth of information that allows analysis to inductively construct the audience for sale. Thus, it takes an unintelligible mass and renders it into easily discernible, legible categories for advertisers to target.
The risks attached to dataveillance are well documented in scholarly literature. Even though databases are often largely constructed with information that is not inherently sensitive about actions that occur publicly, discrimination and damage may still flow from the analysis and application of findings.
To understand this, it is critical not to dissociate data from the people it describes -- "Any data might be sensitive, depending upon the context in which it is used," -- not the context in which it is collected (5. source: Raab and Bennett, 1998, p. 266).
Zarsky provides a useful overview of the social impacts of dataveillance and it is clear that the possible outcomes are never clear-cut. In addition, he suggests that datamining does not precipitate entirely new outcomes so much as augment and expedite processes that already exist.
The two most complicated issues Zarsky examines are the likelihood of discrimination and the threat to autonomy. Of the former, he believes that it is possible to view the effect that datamining tools would have on discrimination based on race or other personal traits "in a different light, and possibly find that the use of such applications should be encouraged" because they might be ameliorative (6. source: Zarsky, 2002, p. 15.).
Not everyone is so sanguine, however, and some see dataveillance and the analysis it leads to as inherently discriminatory particularly for poor people.
Zarsky himself notes that the notion of beneficial price discrimination is threatened by the occurrence of market failures. Regarding the threat to autonomy, Zarsky is concerned. Interestingly, the hazard here develops without vendor malice.
Essentially, as technology is employed to customize and enhance services, diversity decreases and with it liberty is weakened (source: Sunstein, 2001). For both the individual and society, this is clearly a deleterious consequence.
As a company that organizes information at its core, Google has not been sufficiently examined in this context. To begin with, Google maintains server records for every search term used, every Web site visited, and the IP address and browser associated with these actions.
Google's other services are not innocuous either. Gmail, in return for free access to a large amount of server space, scans every e-mail and provides targeted advertisements based on the words within them.
If logged in, the account created to use Gmail carries over to use of Froogle or Google Scholar, linking academic and consumer searches with a unique user regardless of the computer used. To complete the profile, Google also offers orkut, a social networking site where registrants may post pictures, interests, and other personal information.
The extent to which leakage between Google's services exists and whether it leads to profiles is impossible to know -- the services may be totally discrete, there might not be any profiling. It is also possible to view even Gmail's AdWords service as anti-discriminatory.
Rather than aiming advertisements on the basis of presumed demographic knowledge and the attendant stereotypes, AdWords targets wholly (and automatically) on the presence of words in an e-mail -- no matter who the sender or recipient may be.
Even if Google were entirely beneficent, it is also conceivable to imagine a security breach similar to the recent incident involving Equifax (7. source: Carly Suppa, 2004. "Credit agency reports security breach," Computerworld, at http://www.computerworld.com/securitytopics/security/story/0,10801,91319,00.html).
Another feasible outcome is the (unintended) political balkanizing described by Sunstein that may develop through use of personalized, intelligent Google (now in beta). Either way, discourse evaluating the changing information landscape and Google's prominent role within it is necessary.
Photo credit: Leah-Anne Thompson
Trust is a concept that is fundamental and disparate, intuitive and indescribable. Though utterly intangible, some parameters bounding trust can be established. First, the interaction must be voluntary and choice must exist (source: Deutsch, 1962).
Second, trust is only relevant to the extent that uncertainty about the interaction exists. These guidelines excluded instances in which the relationship between actors may be skewed such that fear or faith fills the role of trust.
The decision to participate in some sort of exchange may be the result of painstaking consideration or an unconscious heuristic, but either way the resulting presence of trust provides the framework for the interaction to proceed.
Depending on the relationship and context, the precise character of trust varies. This said, at root, trust may be thought of as the expectations that emerge when a truster's "perception of the trustee's competence [is] defined as the technical ability or expertise to perform the expected behavior," combined with the belief that the trustee will not exact harm (8. source: Duda, p. 48).
In business, trust is a critical component of success. For a consumer to participate in an exchange with a business, she must trust that the goods or services will be rendered as promised and that there will not be negative, unforeseen consequences (source: D. Koehn, 1996. "Should we trust in trust?" American Business Law Journal, volume 34, number 2, pp. 183-203.Koehn, 1996).
To develop this trust, businesses may rely on a host of factors including their physical presence, reputation, membership in trade organizations, or sanctioning from government (source: Kracher, et al., 2005). Ultimately, trust is developed over time and is the accumulation of experience gained through many disparate interactions (Dasgupta, 1988).
In e-commerce, "Trust has been posited as the most important element," of successful exchanges (9. source Kracher, et al., p.131). Trust in an online environment has two important obstacles not found in offline interactions (source: Riegelsberger, et al., 2003).
First, risk is inherently higher in mediated interactions. Intermediation forces the necessity of higher a priori trust because less information is readily available. To compound this problem, it is possible that there are critical differences -- linguistic, cultural, or locational -- between the actors that complicate the exchange.
Second, the interaction is complicated by the intervention of the enabling technologies. Though these technologies are used to facilitate the interaction, they are an additional level of friction that requires attention.
In spite of these obstacles, the speed, ease, and accessibility offered by the application of information technologies are persuasive reasons to use them in commercial transactions. Gaining trust in an online environment is necessary, and while some characteristics of developing trust are similar to those necessary for bricks-and-mortar commerce, there are differences.
In a Cheskin Research (1999) study, six components for cultivating trust online are outlined: seals of approval, brand, navigation, fulfillment, presentation, and technology.
The first three items are self-explanatory; fulfillment refers to the accessibility of information about billing, shipping, and returns; presentation refers to the design interface and the connotation of professionalism; and, technology refers to the presence of cutting-edge technology which evinces professionalism as well.
Parsing trust and e-commerce in this way provides a useful way for understanding how trust develops. It also provides a framework for understanding discourse about an e-commerce company.
Assuming that trust is necessary to commercial success, its value may be best seen when there is reason to believe it has been lost. Three examples follow that help illuminate how media coverage affects trust, and to what consequence.
In June 1999, the proposed merger of DoubleClick and Abacus, two companies engaged in online marketing, touched off a storm of protest from privacy advocates that led to an U.S. Federal Trade Commission (FTC) investigation (source: Macavinta, 2000).
At the time of the merger announcement, DoubleClick shares traded on the Nasdaq for about US$90; by the close of the FTC probe five months later, shares had dropped to nearly US$30 (before recovering after the FTC's report) (10. source: Share prices are drawn from public, historical price indices available at finance.yahoo.com).
In 2002, Microsoft endured an antitrust suit in the United States. Like DoubleClick, Microsoft's shares decreased in value as the story played itself out in the media, falling from about US$36 to US$21 before settlement of the suit.
Annual corporate reputation quotient rankings provided by Harris Interactive may provide a more direct measure of trust in this case: Microsoft had the second-highest rating at the end of 2001 and had fallen to thirteenth at the close of 2002.
In both of these cases, government intervention surely contributed to the company's woes, but the Harris ratings show that media coverage was a component as well.
A final example of the damages that lost trust may cause can be seen in the end of Dan Rather's tenure as host of the CBS Evening News. After basing a report about President Bush on a forged document during the 2004 campaign, CBS and Rather endured a maelstrom of negative publicity and bottomed-out in the ratings.
Shortly thereafter, Rather announced his retirement -- and CBS still lags far behind NBC and ABC in the Nielsen ratings today.
In addition, just five years after acquiring CBS, in June, 2005 Viacom chose to spin CBS (and some related units) off into its own entity partly because the network was seen to be a weight on the corporation's stock price. All three companies deal, to some extent, with information; in all three cases, media coverage certainly precipitated public awareness and contributed to the negative ramifications.
CBS, like Google, bases its business around providing information to the public. Because of this, the difficulties its news division faces after being tarnished in 2004 may be most analogous to the stakes at hand for Google.
End of Part 1
About the author:
Lee Shaker is a Ph.D. candidate at the Annenberg School for Communication, University of Pennsylvania and holds a B.A. in History from University of California, Santa Barbara. His research interests include media use, local politics, and the institutional pressures that shape new media technologies. Please direct comments or questions to lshaker [at] asc [dot] upenn [dot] edu.
1. Battelle, 2005, p. 15.
2. Gandy, 1993, p. 15.
3. Lyon, 2003, p. 20.
4. Arvidsson, 2004, p. 457.
5. Raab and Bennett, 1998, p. 266.
6. Zarsky, 2002, p. 15.
7. Carly Suppa, 2004. "Credit agency reports security breach" Computerworld
8. Duda, 2004, p. 48.
9. Kracher, et al., 2005, p. 131.
10. Share prices are drawn from public, historical price indices available at finance.yahoo.com.
11. D'Angelo, 2002, p. 877; Pan and Kosicki, 1993, p. 58.
12. This total was arrived at in the following manner. First, the primary topic of the article had to be Google for the quotes to be tracked -- this eliminated the 17 articles that mentioned Google only incidentally. Second, press releases or other public information was not counted as a 'unique' source; only individuals that were sought out specifically by the reporter were included. Third, no matter how many separate quotes were in each article, each source was only counted once per article -- though they may have been counted again if they were referred to in another article.
R. Alsop, 2005. "Ranking corporate reputations" Wall Street Journal (6 December), p. B1
A. Arvidsson, 2004. "On the 'pre-history of the panoptic sort": Mobility in market research," Surveillance and Society, volume 1, number 4, pp. 456-474
C.E. Baker, 2002. Media, markets, and democracy. Cambridge: Cambridge University Press
J. Battelle, 2005. The search: How Google and its rivals rewrote the rules of business and transformed our culture. New York: Portfolio
K.M. Carragee and W. Roefs, 2004. "The neglect of power in recent framing research" Journal of Communication, volume 54, number 2, pp. 214-233
Cheskin Research and Studio Archetype/Sapient, 1999. E-commerce trust study
P. D'Angelo, 2002. "News framing as a multiparadigmatic research program: A response to Entman" Journal of Communication, volume 52, number 4, pp. 870-888
A. Danna and O.H. Gandy, Jr., 2002. "All that glitters is not gold" Journal of Business Ethics, volume 40, pp. 373-386
P. Dasgupta, 1988. "Trust as a commodity" In: D. Gambetta (editor). Trust: Making and breaking cooperative relations. New York: Basil Blackwell, pp. 49-72
M. Deutsch, 1962. "Cooperation and trust: Some theoretical notes" In: Nebraska Symposium on Motivation. Lincoln: University of Nebraska Press, pp. 275-318
N.L. Duda, 2004. "The intersection of privacy concern and trust: Beliefs of privacy orientation groups" dissertation, University of Pennsylvania
Elmer, 2004. Profiling machines: Mapping the personal information economy. Cambridge, Mass.: MIT Press
R.M. Entman, 1993. "Framing: Toward clarification of a fractured paradigm" Journal of Communication, volume 43, number 4, pp. 51-58
O.H. Gandy, 1993. The panoptic sort: A political economy of personal information. Boulder, Colo.: Westview
O.H. Gandy, 1982. Beyond agenda setting: Information subsidies and public policy. Norwood, N.J.: Ablex
D. Good, 1988. "Trust as a commodity" In: D. Gambetta (editor). Trust: Making and breaking cooperative relations. New York: Basil Blackwell, pp. 31-48
Harris Interactive, 2005. "Archived Annual Reputational Quotient Results & Reprints"
H.B. Herring, 2004. "In the Internet they trust (your ad here)" New York Times (1 August), Section 3, p. 2
P. Johnson, 2005. "Will CBS News' multiple-anchor concept fly?" USA Today (26 January)
D. Koehn, 1996. "Should we trust in trust?" American Business Law Journal, volume 34, number 2, pp. 183-203
B. Kracher, C.L. Corritore, and S. Wiedenbeck, 2005. "A foundation for understanding online trust in electronic" Information, Communication, & Ethics in Society, volume 3, number 3, pp. 131-141
D. Lyon (editor), 2003. Surveillance as social sorting: Privacy, risk, and digital discrimination. London: Routledge
C. Macavinta, 2000. "Privacy fears raised by DoubleClick database plans" CNET News (25 January)
E. Mills, 2005 "Google balances privacy, reach" CNET News (14 July)
D.C. Mutz, 2005. "Social trust and e-commerce: Experimental evidence for the effects of social trust on individuals' economic behavior" Public Opinion Quarterly, volume 69, number 3, pp. 393-416
H. Nissenbaum, 1998. "Protecting privacy in an information age: The problem of privacy in public" Law and Philosophy, volume 17, pp. 559-596
A. Odlyzko, 2003. "Privacy, economics, and price discrimination on the Internet," N. Sadeh (editor). ICEC2003: Fifth international conference on electronic commerce. New York: ACM Press, pp. 355-366; and at http://www.dtc.umn.edu/~odlyzko/doc/privacy.economics.pdf.
S. Olsen, 2005. "eBay launches national TV ad campaign" ZDNet News (26 October)
A. Orlowski, 2004b. "Google mail is evil -- privacy advocates," Register (3 April)
Z. Pan and G.M. Kosicki, 1993. "Framing analysis: An approach to news discourse," Political Communication, volume 10, pp. 55-75
D.J. Phillips, 2003. "Beyond privacy: Confronting locational surveillance in wireless communication," Communication Law and Policy, volume 8, number 1, pp. 1-23
C.D. Raab and C.J. Bennett, 1998. "The distribution of privacy risks: Who needs protection?" Information Society, volume 14, number 4, pp. 253-262
L. Rainie, 2005. "Search engine use shoots up in the past year and edges towards e-mail as the primary Internet application," Pew Internet and American Life Project
J. Riegelsberger, A.M Sasse, and J.D. McCarthy, 2003. "The researcher's dilemma: Evaluating trust in computer-mediated communication" International Journal of Human-Computer-Studies, volume 58, number 6, pp. 759-781
D.A. Scheufele, 1999. "Framing as a theory of media effects" Journal of Communication, volume 49, number 1, pp. 103-121
C.R. Sunstein, 2001. Republic.com. Princeton, N.J.: Princeton University Press
Carly Suppa, 2004. "Credit agency reports security breach," Computerworld
W. Swarts, 2005. "Viacom's split personality," SmartMoney.com (21 March)
G. Tuchman, 1978. Making news: A study in the construction of reality. New York: Free Press
T.Z. Zarsky, 2002. "'Mine Your Own Business!'" Making the case for the implications of the data mining of personal information in the forum of public opinion" Yale Journal of Law & Technology, volume 5Lee Shaker - [ Read more ]
blog comments powered by Disqus