Spinspotter.com and NewsTrust.net attempt a similar mission of assisting web information gatherers in making judgments on a news source's reliability and objectivity. With the increasing amount of information available from sources outside legacy media, such devices provide an important tool to sift through the fat of the internet. They also give readers an opportunity to make their own assessments of media bias. The user-generated nature of these sites creates some pitfalls and inconsistencies but, far from letting readers off the hook by providing judgment for them, sites like these engage the readers' critical judgment, making them more shrewd information consumers.
Newstrust follows a fairly standard format of news aggregation and
user-generated reviews on MSM articles, independent articles, and MSM
opinion articles. A star rating system appears as you might see on
restaurant reviews on yelp. Users
rate news based on accuracy, balance, context, evidence, fairness,
importance, information, sources, style, and trust. Users are very
familiar with formats like this. It provides a good venue for finding
articles with potential problems and sharing opinions on them. As in
restaurant reviews on yelp, the star rating system is not hard and
fast. Some reviews must be taken with a grain of salt based on each
person's own fairness and bias.
Spinspotter attempts something a little more ambitious, though the usefulness of the features that diverge from the more traditional model is somewhat dubious. Now in its early beta phase, spinspotter installs a toolbar onto web browsers that "spots spin," highlights it in red with markers that open as popups to explain the journalistic violation. At present these markers all seem to be created by individual users but as the system of markers becomes more dense, algorithms based off of these user-generated markers will identify and mark spin on any website visited automatically. That is where the project goes awry.
The sparsely placed markers that I was able to find on nytimes.com and foxnews.com, two sites suggested by the spinspotter home page as good bets for finding spin alarms, illustrated the problem with turning user-generated commentary into an anonymous mathematical system. The tool is pretty neat, scanning each new page and letting users write their comments on markers right on top of the text, rather than having to go to an outside website. It provides an innovative and effective way to give and assess feedback on news articles while you're reading them. The spin markers are based on a system of rules adapted from various journalistic ethics codes, but much like restaurant reviews on yelp, ratings are personal, and can be unfair.
Unsurprisingly, flags on The New York Times decried its liberal bias, while those on Fox News revealed its conservative bias. Some comments sounded like the opinions you would read on a forum or in the comments section below the article but were held up by a specific ethics violation.
In an article
in The NY Times about the upcoming debates, user "bcaro" flags the use
of the word "loquacious" to describe Joe Biden as spin, citing the
"Reporter's Voice" violation. But the description given seems to
suggest that it is not the use of the adjective that upset the user,
but the choice of the adjective he objects to. Bcaro gives a list of
more negative sounding synonyms and concludes "I gues [sic] it was the
'prettiest' sounding word for what he is." As if the description would
have been valid if it had fit the user's opinion of Biden.
In the same article, bcaro flags the use of "Mr. McCain" as it appears right next to Barack Obama's introduction as "Senator Barack Obama of Illinois, the Democratic nominee for president, and Mr. McCain ... " The violation cited in this instance is Disregarded Context because the author failed to refer to McCain as Senator McCain even though Obama is referred to in such a way. The description on this one is highly personal: "umm...the Republican candidate is SENATOR McCain, just like the Democrat candidate is SENATOR Obama. (This was an easy one, folks. Jump in, the water's fine!)" What the user fails to realize is that John McCain was already introduced in the article in the exact same phrase four paragraphs up from the reference flagged. In accordance with New York Times style conventions, the candidate is referred to in full at first mention, and thereafter referred to as Mr. McCain or Obama. Because this was Obama's first introductory mention, he is referred to in full, while McCain gets the shortened referral since he was introduced earlier in the article.
Idiosyncrasies in user flags like these two demonstrate what a farce an algorithm spin identifier would be. If comments like these contribute to the algorithm such that every time McCain is referred to as "Mr. McCain" it is considered spin, the spinoculars are not only useless but unaccountable. At least these flags are attributed to bcaro and readers can see the personal description and make up their own mind. But if an algorithm based on attributable opinions is making decisions without linking to the opinions that informed it, what context do users have to judge the legitimacy of the flags?
An algorithm set to simply identify a pre-established set of charged words within articles might be a useful tool, but an algorithm based on user interpretation of complicated and debated concepts of ethics equals something lesser than the parts of the whole.
Spinspotter attempts something a little more ambitious, though the usefulness of the features that diverge from the more traditional model is somewhat dubious. Now in its early beta phase, spinspotter installs a toolbar onto web browsers that "spots spin," highlights it in red with markers that open as popups to explain the journalistic violation. At present these markers all seem to be created by individual users but as the system of markers becomes more dense, algorithms based off of these user-generated markers will identify and mark spin on any website visited automatically. That is where the project goes awry.
The sparsely placed markers that I was able to find on nytimes.com and foxnews.com, two sites suggested by the spinspotter home page as good bets for finding spin alarms, illustrated the problem with turning user-generated commentary into an anonymous mathematical system. The tool is pretty neat, scanning each new page and letting users write their comments on markers right on top of the text, rather than having to go to an outside website. It provides an innovative and effective way to give and assess feedback on news articles while you're reading them. The spin markers are based on a system of rules adapted from various journalistic ethics codes, but much like restaurant reviews on yelp, ratings are personal, and can be unfair.
Unsurprisingly, flags on The New York Times decried its liberal bias, while those on Fox News revealed its conservative bias. Some comments sounded like the opinions you would read on a forum or in the comments section below the article but were held up by a specific ethics violation.
In an article
in The NY Times about the upcoming debates, user "bcaro" flags the use
of the word "loquacious" to describe Joe Biden as spin, citing the
"Reporter's Voice" violation. But the description given seems to
suggest that it is not the use of the adjective that upset the user,
but the choice of the adjective he objects to. Bcaro gives a list of
more negative sounding synonyms and concludes "I gues [sic] it was the
'prettiest' sounding word for what he is." As if the description would
have been valid if it had fit the user's opinion of Biden.In the same article, bcaro flags the use of "Mr. McCain" as it appears right next to Barack Obama's introduction as "Senator Barack Obama of Illinois, the Democratic nominee for president, and Mr. McCain ... " The violation cited in this instance is Disregarded Context because the author failed to refer to McCain as Senator McCain even though Obama is referred to in such a way. The description on this one is highly personal: "umm...the Republican candidate is SENATOR McCain, just like the Democrat candidate is SENATOR Obama. (This was an easy one, folks. Jump in, the water's fine!)" What the user fails to realize is that John McCain was already introduced in the article in the exact same phrase four paragraphs up from the reference flagged. In accordance with New York Times style conventions, the candidate is referred to in full at first mention, and thereafter referred to as Mr. McCain or Obama. Because this was Obama's first introductory mention, he is referred to in full, while McCain gets the shortened referral since he was introduced earlier in the article.
Idiosyncrasies in user flags like these two demonstrate what a farce an algorithm spin identifier would be. If comments like these contribute to the algorithm such that every time McCain is referred to as "Mr. McCain" it is considered spin, the spinoculars are not only useless but unaccountable. At least these flags are attributed to bcaro and readers can see the personal description and make up their own mind. But if an algorithm based on attributable opinions is making decisions without linking to the opinions that informed it, what context do users have to judge the legitimacy of the flags?
An algorithm set to simply identify a pre-established set of charged words within articles might be a useful tool, but an algorithm based on user interpretation of complicated and debated concepts of ethics equals something lesser than the parts of the whole.
Leave a comment