authors' reply re NYT op-ed "More Is Not Necessarily Better"

[Dave Farber <dave () farber net>]

___


...... Forwarded Message .......
From: "K. N. Cukier" <kn () cukier com>
To: Dave Farber <dave () farber net>
Cc: Matt Hindman <hindman () gmail com>
Date: Thu, 26 Aug 2004 19:57:02 +0100
Subj: authors' reply re NYT op-ed "More Is Not Necessarily Better"

Dave,

We'd be grateful if you'd post this note to the list, as a quick reply 
to the thoughtful criticisms on IP to our NYT op-ed on Google and the 
concentration of online information. Many thanks.

KNC & MH
_______________

Reply to IP List Comments:

If truth is the first casualty of war, then nuance is the first victim 
of journalism.... We appreciate Gerry's and Eszter's comments on IP 
about our op-ed in Monday's New York Times, about the social 
implications of Google. Since there is more to our views than can be 
crammed into 700 words for the mainstream media, we'd like to offer a 
quick response (and added a few cites to our work at the bottom).

Gerry's criticism, in short, is: "No-duh, and it's a good thing, too!" 
Eszter suggests that our stats overstate the case (her other points, 
which focus on users' lack of skill, supports our central argument). 
We'll address each separately, but first state: 1. Google is of course 
a godsend for finding information online, and 2. the phenomenon we 
discuss, that the technology winnows what people see, is not Google's 
fault (though their approach extenuates the situation).

To Gerry's point, we agree that there is an overabundance of 
information, and the use of filters -- be it newspaper editors or new 
technologies -- isn't newsworthy. But if filtering is inevitable, it is 
equally clear that how these filters work matters for society. In the 
case of newspapers, they have inherent and fairly understood biases; 
sometimes, such as with the WSJ, it's worn on their sleeves. Likewise, 
technology is rarely neutral, though why so is less apparent. It is 
probably fair to say that most people are unaware of the tradeoffs 
between locating information online and the diversity of information 
that is visible, which happens every time they use a search engine. The 
article was meant to raise that awareness.

Gerry bases his criticism on the idea that the great mass of overlooked 
information must necessarily be uninteresting anyway. That is a hard 
statement to accept. Consider what the hierarchical nature of the Web 
means on an empirical level, for instance, for a search on a sensitive 
topic like abortion. Google identifies 4,010,000 pages; Matt's research 
(link at bottom) suggests that the top-site in that category accounts 
for roughly 30% of all in-bound links (and thus is most prominent in 
searches), the top 10 pages account for 70% of all links, and the top 
50 almost 90%. Stop to realize what this means: the other 4,009,950 
pages on the topic get only 10% of all in-bound links combined -- and 
as a result, they are essentially invisible online. Is most of this 
content poor? Probably. All of it? Surely not.

As for Eszter's comment that the stats we use overstate our case, we 
disagree. The data that Google's and Yahoo's technology account for 95% 
of all searches comes from comscore Media Metrix, which despite its 
limitations, is nevertheless the best stats available (see: 
http://searchenginewatch.com/reports/article.php/2156431). In the 
article, we use this data to highlight market power, not agreement 
among search results, and in this context the 95% figure is very 
appropriate. (Microsoft's intention to enter the search market will 
bring the number of competitors to three -- but the fact it requires a 
multi-billion-dollar investment to do so actually reinforces our 
point.)

Strikingly, search engine convergence has grown enormously since March 
2004 when Eszter published her First Monday article, which she cited in 
her note 
(http://www.firstmonday.org/issues/issue9_3/hargittai/index.html). Her 
piece explains that though AOL and myway.com are based on Google's 
technology, the results are different. However, that is no longer true. 
Take her example of the query "taxes." Entering the term into 
search.aol.com or myway.com now returns EXACTLY the same results as 
Google itself. Try it. Moreover, Yahoo's results for "taxes" are now 
quite close to Google's: the same top site, with seven of the top ten 
now identical.

Taken together, these examples show that the use of link-structure to 
identify content effectively limits the diversity of what the average 
user is exposed to online, even if one uses multiple search engines. 
Compared to the bad old days of search, this seems a tradeoff worth 
making. At the same time, though, we shouldn't be blind to the social 
consequences of search engine convergence.

We're thankful for all the feedback we've received. On a final note, 
the NYT piece served as our swan-song at the NCDG. After two wonderful 
and productive years, we've both recently left our fellowships at the 
Kennedy School; Matt to teach politics at Arizona State University, and 
myself, to cover technology and policy issues for The Economist in 
London. We both look forward to discussing these issues more going 
forward.

Best,

Kenn & Matt
______________________

Some references to our previous work on the topic (going from somewhat 
simplified to more complex):

"The Web, Media Diversity and Power Laws"
Harvard University NCDG Resource Page
http://www.ksg.harvard.edu/digitalcenter/powerlaw.htm

"More Media, Less Diversity"
The New York Times, op-ed page, June 2, 2003
http://www.princeton.edu/~mhindman/NYTimesOpEd2.htm

"Measuring Media Concentration Online and Offline"
Ford Foundation conference on media diversity, Dec. 2003.
http://www.cukier.com/writings/webmedia-jan04.htm

"Googlearchy: How a Few Heavily-Linked Sites Dominate Politics Online"
Early draft chapter of Matt's dissertation
http://www.princeton.edu/~mhindman/googlearchy--hindman.pdf

To contact the authors:
Matthew Hindman -- www.hindman.cc -- hindman () gmail com
Kenneth Neil Cukier -- www.cukier.com -- cukier () gmail com
(Please note our gmail accounts -- See! We are not anti-Googlists!)

END


-------------------------------------
You are subscribed as interesting-people () lists elistx com
To manage your subscription, go to
  http://v2.listbox.com/member/?listname=ip

Archives at: http://www.interesting-people.org/archives/interesting-people/