3 ms·

A different or supplementary approach could be removing the bias from a news item. Instead of trying to determine which biased side an article is skewed toward

by qwrusz 10y ago

A different or supplementary approach could be removing the bias from a news item. Instead of trying to determine which biased side an article is skewed towards and then finding other links to what is determined to be the "alternative view". Scrubbing bias or at least highlighting it is already helpful. For example: at work we get daily emailed briefs with major business news items summarized to ~3-5 bullet points of facts. Journalist opinions/bias and rhetoric language is mostly removed in the bullet point sentences. It's not a perfect system by any means, not even close, and I would love to see something similar offered that's improved and expanded in what it can do. This type of bias scrubbing/summarizing is easier in business news and sports news which involve more numbers and figures reporting (+nowadays many of the full articles may also be written entirely by bots - see link below). It would be harder to expand this for longer investigative/politics news articles. But a partial imperfect solution here is better than status quo. I would be a user of a tool that could summarize key "unbiased facts" from articles and I would be interested in helping build it too. Link to a NYT story about algos writing/summarizing news: http://www.nytimes.com/2015/03/08/opinion/sunday/if-an-algorithm-wrote-this-how-would-you-even-know.html?_r=0 http://www.nytimes.com/2015/03/08/opinion/sunday/if-an-algor...

hga 10y ago
What if the bias was created by omitting important facts that, for example, contradict the narrative of the item. Sports are indeed a good domain for what you're talking about, but that's because they're run by well known rules, it's a highly artificial domain, not directly including the messiness of general human affairs (except in that they're played, refereed, managed etc. by humans). Business ... well, how many stories about Yahoo! have ignored its negative market net worth when its Alibaba stake is removed, when that was relevant? Not necessarily examples of bias, but....
- qwrusz 10y ago
  Good point about omission of important facts. It's tough because facts can be absent from a news item intentionally and unintentionally. Journalists have time and size constraints - so even when striving for accuracy and objectivity, facts may just be missed and some other known facts will certainly be left out as determined not relevant enough at the time of writing. This is just the reality of reporting news. Professional Journalism has ethics and standards - of course w/ ongoing debates in how they are practiced - and they are taken seriously by the industry as a whole. Wikipedia has more info: https://en.wikipedia.org/wiki/Journalism_ethics_and_standards https://en.wikipedia.org/wiki/Journalism_ethics_and_standard... Ultimately, for news items where a reader wants to get a fuller picture of the story he/she is going to have to go read multiple longer articles about it. Scrubbed bullet points may offer discovery with less bias at this initial stage and from which to choose those stories to go read more about. I don't want to get off topic, but the Yahoo example mentioned is interesting and I think worth maybe discussing so TL;DR skip this next part if very wonky Yahoo stock news reporting is not an interest: Where did you get your Yahoo business news?...Firstly, Yahoo's (YHOO) value is not negative ex Alibaba (BABA). "A company's shares are worth what someone is willing to pay for them" is a relevant quote here and Verizon offered ~$4.8 Billion to buy Yahoo ex Alibaba (also ex Yahoo Japan btw). Pretty big difference between a negative value and $4.8Billion! Second, if talking about a stub Yahoo valuation? Like using some shorthand value estimation method based on looking at: (BABA's recent stock prices)x(size of Yahoo's stake) and comparing that amount to where YHOO's market cap is trading at? Then yes you could end up at all kinds of crazy totals, including negative numbers, depending on what shorthand valuation method you chose to use... But these types of calculations, maybe interesting and fun, are basically meaningless in the real world. The amount Yahoo will receive for its Alibaba stake at some unknown time in the future cannot be calculated accurately from BABA's stock price today. Nor can it derived from whatever $ amount Yahoo has written in its books for the value of their BABA stake (Book value <> market value). (Sum of parts also <> market value) Yahoo can be trading at a discount (or a premium) to one's hypothetical future value calculations for all kinds of reasons: eg YHOO's potential +$10B tax liability uncertainty should they sell their BABA shares with capital gains (Yahoo still talking to IRS about this and its unclear how it will play out but it's shocking Yahoo has not been given clearer answers by the IRS yet to quite fair questions about the tax code) Yahoo is also still facing material issues in Verizon deal not yet finalized (YHOO email breach may change terms, this is also unclear), etc.. etc.. Basically, Stock Valuation Is Not News. Though some stock trading websites look like news or are attached to legit news sites (Barron's to WSJ for example). Barron's is not news. NYTimes DealBook discussed Yahoo stub being negative. DealBook is a blog. http://www.nytimes.com/2015/12/10/business/dealbook/how-to-value-yahoos-core-business.html?_r=0 http://www.nytimes.com/2015/12/10/business/dealbook/how-to-v... An analogy for those not familiar with Yahoo's stock price valuation: This is like guessing how much money you could receive from selling your house if you tore it down and sold it for scrap wood and separately put the plot of land on the market. You also have to calculate this future value without knowing if your local council will approve the idea, or how long it will take, or what prices for scrap wood/plots of land will be, or how much tearing down your house will cost, or the % odds you secretly have been living in a house built over an old unicorn graveyard and can't sell the land at all. One can imagine all these factors are why a number calculated for future "value" of stock may be very different than what actually is received when a sale happens. Like way off. Like a negative number vs. +$4.8Billion way off.

by qwrusz 10y ago

hga 10y ago

What if the bias was created by omitting important facts that, for example, contradict the narrative of the item. Sports are indeed a good domain for what you're talking about, but that's because they're run by well known rules, it's a highly artificial domain, not directly including the messiness of general human affairs (except in that they're played, refereed, managed etc. by humans). Business ... well, how many stories about Yahoo! have ignored its negative market net worth when its Alibaba stake is removed, when that was relevant? Not necessarily examples of bias, but....

qwrusz 10y ago
Good point about omission of important facts. It's tough because facts can be absent from a news item intentionally and unintentionally. Journalists have time and size constraints - so even when striving for accuracy and objectivity, facts may just be missed and some other known facts will certainly be left out as determined not relevant enough at the time of writing. This is just the reality of reporting news. Professional Journalism has ethics and standards - of course w/ ongoing debates in how they are practiced - and they are taken seriously by the industry as a whole. Wikipedia has more info: https://en.wikipedia.org/wiki/Journalism_ethics_and_standards https://en.wikipedia.org/wiki/Journalism_ethics_and_standard... Ultimately, for news items where a reader wants to get a fuller picture of the story he/she is going to have to go read multiple longer articles about it. Scrubbed bullet points may offer discovery with less bias at this initial stage and from which to choose those stories to go read more about. I don't want to get off topic, but the Yahoo example mentioned is interesting and I think worth maybe discussing so TL;DR skip this next part if very wonky Yahoo stock news reporting is not an interest: Where did you get your Yahoo business news?...Firstly, Yahoo's (YHOO) value is not negative ex Alibaba (BABA). "A company's shares are worth what someone is willing to pay for them" is a relevant quote here and Verizon offered ~$4.8 Billion to buy Yahoo ex Alibaba (also ex Yahoo Japan btw). Pretty big difference between a negative value and $4.8Billion! Second, if talking about a stub Yahoo valuation? Like using some shorthand value estimation method based on looking at: (BABA's recent stock prices)x(size of Yahoo's stake) and comparing that amount to where YHOO's market cap is trading at? Then yes you could end up at all kinds of crazy totals, including negative numbers, depending on what shorthand valuation method you chose to use... But these types of calculations, maybe interesting and fun, are basically meaningless in the real world. The amount Yahoo will receive for its Alibaba stake at some unknown time in the future cannot be calculated accurately from BABA's stock price today. Nor can it derived from whatever $ amount Yahoo has written in its books for the value of their BABA stake (Book value <> market value). (Sum of parts also <> market value) Yahoo can be trading at a discount (or a premium) to one's hypothetical future value calculations for all kinds of reasons: eg YHOO's potential +$10B tax liability uncertainty should they sell their BABA shares with capital gains (Yahoo still talking to IRS about this and its unclear how it will play out but it's shocking Yahoo has not been given clearer answers by the IRS yet to quite fair questions about the tax code) Yahoo is also still facing material issues in Verizon deal not yet finalized (YHOO email breach may change terms, this is also unclear), etc.. etc.. Basically, Stock Valuation Is Not News. Though some stock trading websites look like news or are attached to legit news sites (Barron's to WSJ for example). Barron's is not news. NYTimes DealBook discussed Yahoo stub being negative. DealBook is a blog. http://www.nytimes.com/2015/12/10/business/dealbook/how-to-value-yahoos-core-business.html?_r=0 http://www.nytimes.com/2015/12/10/business/dealbook/how-to-v... An analogy for those not familiar with Yahoo's stock price valuation: This is like guessing how much money you could receive from selling your house if you tore it down and sold it for scrap wood and separately put the plot of land on the market. You also have to calculate this future value without knowing if your local council will approve the idea, or how long it will take, or what prices for scrap wood/plots of land will be, or how much tearing down your house will cost, or the % odds you secretly have been living in a house built over an old unicorn graveyard and can't sell the land at all. One can imagine all these factors are why a number calculated for future "value" of stock may be very different than what actually is received when a sale happens. Like way off. Like a negative number vs. +$4.8Billion way off.