Connect with us

Government

Google Denies “Political Bias” In Search Algo Despite Collapse In Conservative Site Visibility

Google Denies "Political Bias" In Search Algo Despite Collapse In Conservative Site Visibility

Published

on

Google Denies "Political Bias" In Search Algo Despite Collapse In Conservative Site Visibility Tyler Durden Tue, 09/22/2020 - 17:25

Authored by Maxim Lott via RealClearPolitics.com,

It has long been feared that Google, which controls almost 90% of U.S. Internet search traffic, could sway an election by altering the search results it shows users. 

New data indicate that may be happening, as conservative news sites including Breitbart, the Daily Caller, and the Federalist have seen their Google search listings dramatically reduced. 

The data come from the search consultancy Sistrix, which tracks a million different Google search keywords and keeps track of how highly different sites rank across all the search terms. 

The tracker shows that Google search visibility for Breitbart first plunged in 2017, before falling to approximately zero in July 2019:

Googling the name "Breitbart" still pulls up the website, but it is nearly eliminated from any searches that don't explicitly name it. For example, Googling the names of Breitbart's reporters sometimes forces users to click through page after page of less-relevant results before hitting a Breitbart link. In the case of Joel Pollak, the first Breitbart link appears on the bottom of page 7 of Google search results. In comparison, a search on the small Google competitor DuckDuckGo gives multiple links to Pollak’s Breitbart work on the first page. 

Breitbart saw that stark reduction in search, even as little else in the news outlet’s reporting model changed. Other conservative news sites, such as the Daily Caller, also were de-ranked at similar times.

Both Breitbart and the Daily Caller have confirmed that their Google traffic fell dramatically as their search rankings fell. 

Internal Google files have hinted at such action. In 2018, the Daily Caller reported on a leaked exchange the day after President Trump's 2016 election win. In it, employees debated whether Breitbart and the Daily Caller should be buried. 

“This was an election of false equivalencies, and Google, sadly, had a hand in it,” Google engineer Scott Byer wrote, according to the documents obtained by the Daily Caller. 

 “How many times did you see …  items from opinion blogs (Breitbart, Daily Caller) elevated next to legitimate news organizations? That’s something that can and should be fixed.” Byer wrote. 

He added, "I think we have a responsibility to expose the quality and truthfulness of sources." 

In the leaked exchange, other Googlers pushed back, saying that it shouldn't be Google's role to rank sources’ legitimacy -- and that hiding sites would only add to distrust. 

Google did not respond to a request for comment about the data presented here.

An expert in search engine optimization pointed RealClearPolitics to a public Google document from 2019 outlining how the company now employs humans to go through webpages and rate them based on "Expertise/Authoritativeness/Trustworthiness." "Google has acknowledged they use human search quality raters who help evaluate search results," said Chris Rodgers, CEO and founder of Colorado SEO Pros.  

Google does not directly use such ratings to rank sites, but "based on those ratings Google will then tweak their algorithm and use machine learning to help dial in the desired results," Rodgers explained. 

The Google guidelines instruct raters to give the "lowest" ranking to any news-related "content that contradicts well-established expert consensus." And how does one determine "expert consensus"? The Google guidelines repeatedly advise raters to consult Wikipedia, which it mentions 56 times: 

"See if there is a Wikipedia article or news article from a well-known news site. Wikipedia can be a good source of information about companies, organizations, and content creators." 

The document cites the Christian Science Monitor as an example.

 "Notice the highlighted section in the Wikipedia article about The Christian Science Monitor newspaper, which tells us that the newspaper has won seven Pulitzer Prize awards," the document tells raters. "From this information, we can infer that the csmonitor.com website has a positive reputation." 

The reliance on Wikipedia could partly explain the de-rankings, as the crowdsourced encyclopedia calls Breitbart "far right" and alleges that the Daily Caller "frequently published false stories." But Wikipedia's co-founder, Larry Sanger, recently wrote an essay about how "Wikipedia is badly biased." 

In addition to the human ratings used to test algorithms, Google also has human-maintained blacklists -- but they are supposed to be very limited. In congressional testimony this summer, Rep. Matt Gaetz asked Google CEO Sundar Pichai if the search engine manually de-ranked websites. Gaetz cited Breitbart, the Daily Caller, as well as WesternJournal.com and Spectator.org, two other conservative news sites that Sistrix data show have been nearly removed from Google searches. 

Pichai responded that his company only removes sites if they are deemed to be "interfering in elections" or conveying "violent extremism." 

It remains unclear why the conservative sites were de-ranked. "Only Google can truly know why these sites have seen ranking losses," SEO expert Rodgers said. 

Jeremy Rivera, of JeremyRiveraSEO.com, told RealClearPolitics that there are several possible reasons for de-ranking. Both Breitbart and the Daily Caller have “low-quality” sites linking to them, he said, and they host many ads, which create slow page-load times. "Page-load speed is a direct ranking factor," he said. 

Several news outlets, on both the left and right, have avoided the near-total blackout experienced by Breitbart and the Daily Caller, but saw their search listings reduced significantly. Among them is The Federalist, which saw search listings fall dramatically this spring after coming under fire for running articles that critics said downplayed the danger of COVID-19.

Some left-leaning news sites were docked too. 

Other major sites, such as the Wall Street Journal, the Washington Post, and CNN, have not seen dramatic changes in rankings, according to Sistrix data. 

Some observers say the trends show innovation at work, as companies try to present users with the highest-quality information. "These are inherently subjective questions," said Berin Szóka of TechFreedom, noting that the First Amendment gives "complete discretion of private media companies, which includes Google. ... They have the same right to decide what content to carry that Breitbart and Fox do." 

Others worry that the near-blacklisting of conservative outlets could tip election outcomes this year. A 2015 peer-reviewed study in the Proceedings of the National Academy of Sciences estimated that a search engine could sway more than 10% of undecided voters in an election simply by altering what results are shown. 

“Such manipulations are difficult to detect, and most people are relatively powerless when trying to resist sources of influence they cannot see," the authors warn. "When people are unaware they are being manipulated, they tend to believe they have adopted their new thinking voluntarily.”

The paper further suggests that Google's 87% market share is a concern. While competitors are growing in popularity, their market share numbers remain low: Microsoft's Bing has 7.2% of the pie, and DuckDuckGo has 1.75%.

“Because the majority of people in most democracies use a search engine provided by just one company... election-related search rankings could pose a significant threat to the democratic system of government,” the paper concludes.

* * *

After publication, a Google spokesperson responded:

"There is no validity whatsoever to these allegations of political bias. Our systems do not take political ideology into account, and we go to extraordinary lengths to build our products for everyone in an apolitical way. Anyone can easily cherry-pick a range of conservative, progressive or non-political sites that have seen traffic changes over time. The improvements we make to Search undergo a rigorous testing process and are done to provide helpful information for the billions of queries we get every day.”

The spokesperson has not responded to specific follow-up questions about whether any progressive news site had seen as stark a fall as Breitbart or the Daily Caller, or whether Google's quality ratings (which are part of their testing process) might be ideologically skewed due to reliance on Wikipedia. 

Read More

Continue Reading

Government

Are Voters Recoiling Against Disorder?

Are Voters Recoiling Against Disorder?

Authored by Michael Barone via The Epoch Times (emphasis ours),

The headlines coming out of the Super…

Published

on

Are Voters Recoiling Against Disorder?

Authored by Michael Barone via The Epoch Times (emphasis ours),

The headlines coming out of the Super Tuesday primaries have got it right. Barring cataclysmic changes, Donald Trump and Joe Biden will be the Republican and Democratic nominees for president in 2024.

(Left) President Joe Biden delivers remarks on canceling student debt at Culver City Julian Dixon Library in Culver City, Calif., on Feb. 21, 2024. (Right) Republican presidential candidate and former U.S. President Donald Trump stands on stage during a campaign event at Big League Dreams Las Vegas in Las Vegas, Nev., on Jan. 27, 2024. (Mario Tama/Getty Images; David Becker/Getty Images)

With Nikki Haley’s withdrawal, there will be no more significantly contested primaries or caucuses—the earliest both parties’ races have been over since something like the current primary-dominated system was put in place in 1972.

The primary results have spotlighted some of both nominees’ weaknesses.

Donald Trump lost high-income, high-educated constituencies, including the entire metro area—aka the Swamp. Many but by no means all Haley votes there were cast by Biden Democrats. Mr. Trump can’t afford to lose too many of the others in target states like Pennsylvania and Michigan.

Majorities and large minorities of voters in overwhelmingly Latino counties in Texas’s Rio Grande Valley and some in Houston voted against Joe Biden, and even more against Senate nominee Rep. Colin Allred (D-Texas).

Returns from Hispanic precincts in New Hampshire and Massachusetts show the same thing. Mr. Biden can’t afford to lose too many Latino votes in target states like Arizona and Georgia.

When Mr. Trump rode down that escalator in 2015, commentators assumed he’d repel Latinos. Instead, Latino voters nationally, and especially the closest eyewitnesses of Biden’s open-border policy, have been trending heavily Republican.

High-income liberal Democrats may sport lawn signs proclaiming, “In this house, we believe ... no human is illegal.” The logical consequence of that belief is an open border. But modest-income folks in border counties know that flows of illegal immigrants result in disorder, disease, and crime.

There is plenty of impatience with increased disorder in election returns below the presidential level. Consider Los Angeles County, America’s largest county, with nearly 10 million people, more people than 40 of the 50 states. It voted 71 percent for Mr. Biden in 2020.

Current returns show county District Attorney George Gascon winning only 21 percent of the vote in the nonpartisan primary. He’ll apparently face Republican Nathan Hochman, a critic of his liberal policies, in November.

Gascon, elected after the May 2020 death of counterfeit-passing suspect George Floyd in Minneapolis, is one of many county prosecutors supported by billionaire George Soros. His policies include not charging juveniles as adults, not seeking higher penalties for gang membership or use of firearms, and bringing fewer misdemeanor cases.

The predictable result has been increased car thefts, burglaries, and personal robberies. Some 120 assistant district attorneys have left the office, and there’s a backlog of 10,000 unprosecuted cases.

More than a dozen other Soros-backed and similarly liberal prosecutors have faced strong opposition or have left office.

St. Louis prosecutor Kim Gardner resigned last May amid lawsuits seeking her removal, Milwaukee’s John Chisholm retired in January, and Baltimore’s Marilyn Mosby was defeated in July 2022 and convicted of perjury in September 2023. Last November, Loudoun County, Virginia, voters (62 percent Biden) ousted liberal Buta Biberaj, who declined to prosecute a transgender student for assault, and in June 2022 voters in San Francisco (85 percent Biden) recalled famed radical Chesa Boudin.

Similarly, this Tuesday, voters in San Francisco passed ballot measures strengthening police powers and requiring treatment of drug-addicted welfare recipients.

In retrospect, it appears the Floyd video, appearing after three months of COVID-19 confinement, sparked a frenzied, even crazed reaction, especially among the highly educated and articulate. One fatal incident was seen as proof that America’s “systemic racism” was worse than ever and that police forces should be defunded and perhaps abolished.

2020 was “the year America went crazy,” I wrote in January 2021, a year in which police funding was actually cut by Democrats in New York, Los Angeles, San Francisco, Seattle, and Denver. A year in which young New York Times (NYT) staffers claimed they were endangered by the publication of Sen. Tom Cotton’s (R-Ark.) opinion article advocating calling in military forces if necessary to stop rioting, as had been done in Detroit in 1967 and Los Angeles in 1992. A craven NYT publisher even fired the editorial page editor for running the article.

Evidence of visible and tangible discontent with increasing violence and its consequences—barren and locked shelves in Manhattan chain drugstores, skyrocketing carjackings in Washington, D.C.—is as unmistakable in polls and election results as it is in daily life in large metropolitan areas. Maybe 2024 will turn out to be the year even liberal America stopped acting crazy.

Chaos and disorder work against incumbents, as they did in 1968 when Democrats saw their party’s popular vote fall from 61 percent to 43 percent.

Views expressed in this article are opinions of the author and do not necessarily reflect the views of The Epoch Times or ZeroHedge.

Tyler Durden Sat, 03/09/2024 - 23:20

Read More

Continue Reading

Government

Veterans Affairs Kept COVID-19 Vaccine Mandate In Place Without Evidence

Veterans Affairs Kept COVID-19 Vaccine Mandate In Place Without Evidence

Authored by Zachary Stieber via The Epoch Times (emphasis ours),

The…

Published

on

Veterans Affairs Kept COVID-19 Vaccine Mandate In Place Without Evidence

Authored by Zachary Stieber via The Epoch Times (emphasis ours),

The U.S. Department of Veterans Affairs (VA) reviewed no data when deciding in 2023 to keep its COVID-19 vaccine mandate in place.

Doses of a COVID-19 vaccine in Washington in a file image. (Jacquelyn Martin/Pool/AFP via Getty Images)

VA Secretary Denis McDonough said on May 1, 2023, that the end of many other federal mandates “will not impact current policies at the Department of Veterans Affairs.”

He said the mandate was remaining for VA health care personnel “to ensure the safety of veterans and our colleagues.”

Mr. McDonough did not cite any studies or other data. A VA spokesperson declined to provide any data that was reviewed when deciding not to rescind the mandate. The Epoch Times submitted a Freedom of Information Act for “all documents outlining which data was relied upon when establishing the mandate when deciding to keep the mandate in place.”

The agency searched for such data and did not find any.

The VA does not even attempt to justify its policies with science, because it can’t,” Leslie Manookian, president and founder of the Health Freedom Defense Fund, told The Epoch Times.

“The VA just trusts that the process and cost of challenging its unfounded policies is so onerous, most people are dissuaded from even trying,” she added.

The VA’s mandate remains in place to this day.

The VA’s website claims that vaccines “help protect you from getting severe illness” and “offer good protection against most COVID-19 variants,” pointing in part to observational data from the U.S. Centers for Disease Control and Prevention (CDC) that estimate the vaccines provide poor protection against symptomatic infection and transient shielding against hospitalization.

There have also been increasing concerns among outside scientists about confirmed side effects like heart inflammation—the VA hid a safety signal it detected for the inflammation—and possible side effects such as tinnitus, which shift the benefit-risk calculus.

President Joe Biden imposed a slate of COVID-19 vaccine mandates in 2021. The VA was the first federal agency to implement a mandate.

President Biden rescinded the mandates in May 2023, citing a drop in COVID-19 cases and hospitalizations. His administration maintains the choice to require vaccines was the right one and saved lives.

“Our administration’s vaccination requirements helped ensure the safety of workers in critical workforces including those in the healthcare and education sectors, protecting themselves and the populations they serve, and strengthening their ability to provide services without disruptions to operations,” the White House said.

Some experts said requiring vaccination meant many younger people were forced to get a vaccine despite the risks potentially outweighing the benefits, leaving fewer doses for older adults.

By mandating the vaccines to younger people and those with natural immunity from having had COVID, older people in the U.S. and other countries did not have access to them, and many people might have died because of that,” Martin Kulldorff, a professor of medicine on leave from Harvard Medical School, told The Epoch Times previously.

The VA was one of just a handful of agencies to keep its mandate in place following the removal of many federal mandates.

“At this time, the vaccine requirement will remain in effect for VA health care personnel, including VA psychologists, pharmacists, social workers, nursing assistants, physical therapists, respiratory therapists, peer specialists, medical support assistants, engineers, housekeepers, and other clinical, administrative, and infrastructure support employees,” Mr. McDonough wrote to VA employees at the time.

This also includes VA volunteers and contractors. Effectively, this means that any Veterans Health Administration (VHA) employee, volunteer, or contractor who works in VHA facilities, visits VHA facilities, or provides direct care to those we serve will still be subject to the vaccine requirement at this time,” he said. “We continue to monitor and discuss this requirement, and we will provide more information about the vaccination requirements for VA health care employees soon. As always, we will process requests for vaccination exceptions in accordance with applicable laws, regulations, and policies.”

The version of the shots cleared in the fall of 2022, and available through the fall of 2023, did not have any clinical trial data supporting them.

A new version was approved in the fall of 2023 because there were indications that the shots not only offered temporary protection but also that the level of protection was lower than what was observed during earlier stages of the pandemic.

Ms. Manookian, whose group has challenged several of the federal mandates, said that the mandate “illustrates the dangers of the administrative state and how these federal agencies have become a law unto themselves.”

Tyler Durden Sat, 03/09/2024 - 22:10

Read More

Continue Reading

Government

Low Iron Levels In Blood Could Trigger Long COVID: Study

Low Iron Levels In Blood Could Trigger Long COVID: Study

Authored by Amie Dahnke via The Epoch Times (emphasis ours),

People with inadequate…

Published

on

Low Iron Levels In Blood Could Trigger Long COVID: Study

Authored by Amie Dahnke via The Epoch Times (emphasis ours),

People with inadequate iron levels in their blood due to a COVID-19 infection could be at greater risk of long COVID.

(Shutterstock)

A new study indicates that problems with iron levels in the bloodstream likely trigger chronic inflammation and other conditions associated with the post-COVID phenomenon. The findings, published on March 1 in Nature Immunology, could offer new ways to treat or prevent the condition.

Long COVID Patients Have Low Iron Levels

Researchers at the University of Cambridge pinpointed low iron as a potential link to long-COVID symptoms thanks to a study they initiated shortly after the start of the pandemic. They recruited people who tested positive for the virus to provide blood samples for analysis over a year, which allowed the researchers to look for post-infection changes in the blood. The researchers looked at 214 samples and found that 45 percent of patients reported symptoms of long COVID that lasted between three and 10 months.

In analyzing the blood samples, the research team noticed that people experiencing long COVID had low iron levels, contributing to anemia and low red blood cell production, just two weeks after they were diagnosed with COVID-19. This was true for patients regardless of age, sex, or the initial severity of their infection.

According to one of the study co-authors, the removal of iron from the bloodstream is a natural process and defense mechanism of the body.

But it can jeopardize a person’s recovery.

When the body has an infection, it responds by removing iron from the bloodstream. This protects us from potentially lethal bacteria that capture the iron in the bloodstream and grow rapidly. It’s an evolutionary response that redistributes iron in the body, and the blood plasma becomes an iron desert,” University of Oxford professor Hal Drakesmith said in a press release. “However, if this goes on for a long time, there is less iron for red blood cells, so oxygen is transported less efficiently affecting metabolism and energy production, and for white blood cells, which need iron to work properly. The protective mechanism ends up becoming a problem.”

The research team believes that consistently low iron levels could explain why individuals with long COVID continue to experience fatigue and difficulty exercising. As such, the researchers suggested iron supplementation to help regulate and prevent the often debilitating symptoms associated with long COVID.

It isn’t necessarily the case that individuals don’t have enough iron in their body, it’s just that it’s trapped in the wrong place,” Aimee Hanson, a postdoctoral researcher at the University of Cambridge who worked on the study, said in the press release. “What we need is a way to remobilize the iron and pull it back into the bloodstream, where it becomes more useful to the red blood cells.”

The research team pointed out that iron supplementation isn’t always straightforward. Achieving the right level of iron varies from person to person. Too much iron can cause stomach issues, ranging from constipation, nausea, and abdominal pain to gastritis and gastric lesions.

1 in 5 Still Affected by Long COVID

COVID-19 has affected nearly 40 percent of Americans, with one in five of those still suffering from symptoms of long COVID, according to the U.S. Centers for Disease Control and Prevention (CDC). Long COVID is marked by health issues that continue at least four weeks after an individual was initially diagnosed with COVID-19. Symptoms can last for days, weeks, months, or years and may include fatigue, cough or chest pain, headache, brain fog, depression or anxiety, digestive issues, and joint or muscle pain.

Tyler Durden Sat, 03/09/2024 - 12:50

Read More

Continue Reading

Trending