Google Showing Election Bias?

[J_RaD]

---------- DO NOT BUST INTO SOME POLITICAL RANT--------

Evidence is mounting more and more each day

www.regated.com/2016/06/google-censorship-trump/

[youtube.com...]

[ibtimes.co.uk...]

Seem google has its hands in the political cookie jar and is skewing search results in order to sway votes.
Or let me rephrase, A global company is working with the US gov to skew online news and search results in order to get their selected candidate elected.

If this isn't a sign of deep corruption, yikes!

and just today, they are suppressing youtube comments on the obama endorsment

[youtube.com...]


Please keep it about THIS and not who you like or don't like.

No matter who you are voting for THIS SHOULD NOT BE GOING ON AT ALL!

[bakedjake]

and let me add in a MOD had to approve this before it was even posted.

(WebmasterWorld hat on) Please keep moderation discussion private per TOS, please and thanks. I'm advocating for keeping this discussion open if we are civil and follow TOS but we are all on notice (including me as a participant) from the moderation team. I don't want to edit the post directly given the sensitivity of what we're talking about. (WebmasterWorld hat off)

I'm just waiting for everyone to drop the defense. Its not like I posted this with a loaded gun...but that didn't stop me from getting surrounded by them.

Nah, it's not personal. I just disagree with the premise. I just don't think Google is that stupid and I do think that the video evidence proffered wasn't collected in any meaningful way.

I wouldn't even know how to test this if I wanted to. Where do we start? Is autocomplete query data sharded by region? Datacenter? Individual datastore? Does the query type or the results of the query matter when querying the autocomplete data? Velocity of query somehow? Broken out word by word by the text parser? Are your queries stored separately from "generic" autocomplete? How many "states" can a user be in on Google (untracked, tracked/not logged in, tracked/logged in) and how does that affect autocomplete? How do other signals of user intent (such as previous sites visited) affect autocomplete? Andy's better at this and can probably think of 100 things that would affect something like this. We'd have to isolate every single one of them.

How do we prove that a technology is doing something when we don't understand the technology?

[edited by: bakedjake at 4:08 pm (utc) on Jun 11, 2016]

[martinibuster]

Wow j_rad, are you still talking about this? I posted a link that debunked SourceFed. I thought after you read the debunking you'd get a clue that it was nonsense. The allegations were false, the link demonstrated it. Now it's on Snopes. Snopes has debunked the claims [snopes.com], too.

The claims are false, period. To insist on this any further, when it's been thoroughly debunked is getting into chemtrail territory. Do you really want to be that person?

[NickMNS]

I think there is a lesson to be learned here that has nothing to do with politics and everything to do with SEO and web-search.

First let me premise my statement by saying that this all bunch of hog-wash, I think Andy and others have done a great job of showing that. Feel free to disagree...

When we observe our environment we have natural tendency to see patterns and as a result of observing patterns we attribute meaning and causation to those patterns. We often see patterns where there are none. A typical example of this is the Hot-Hand Fallacy. Or we observe patterns and observe some correlation with those patterns and thus attribute causation. Like when the Romans looked to the night skies and watched the stars rotate around the world, and concluded that world was the center of the universe.

The beauty of this story is that is so obviously untrue, yet when observed as it was presented, it appears to be undeniably true. We face many other situations in our field, such as apparent algo updates, SEO signals, and traffic patterns that often are presented in a similar fashion. To assign causation, because of x, then y, in any meaningful way is an extremely difficult thing to do, some may say nearly impossible.

The Google search algo is extremely complex, it is unimaginable that it would be possible even for Google to influence it for a single search term (or small subset of terms) without it having significant negative, overseen and uncontrollable impacts on wide array of other terms. I doubt that Google itself would be able to predict what the auto complete results would be a for any specific search term. This is not to say that they have no control over how auto-complete terms are selected, rather the control is on the aggregate level, across wide array of results.

[J_RaD]

10-4
----
I didn't think google was that stupid either.... until they have been caught changing things. It seems they have a team of trained monkey's searching the web..and when things get pointed out and spread on twitter and facebook, they make edits... not even going to go down that hole but i watched on the sidelines as this happened 5 or 6 times *shaking my head* to say the least.

Which actually led to this post.. after watching that go down 5 or 6 times in over the course of a month... this is not even a stretch to me. I understand if this seems kinda out of the blue
to the avg person... but goog has been making some fishy changes.

Maybe its not even them being stupid? maybe they are doing it because they KNOW they can get away with it? I mean look at us... we are having a convo of how even to prove it.....?!?

auto complete is just one part, search results are another.. the results can be tested the skewed auto complete is just the 1st layer on the cake

*promise I don't ware tin foil hats* :-)

[tangor]

Google is not stupid. They just provide different (counter when political) results than every other search engine. And simply because they won't allow a bad term next to a person's name? This is called plausible deniability .... but it does have the singular effect of REMOVING contrary reporting from the serps.

Despite all the how complex and large the g algo might be commentary, the tweak that does this is so freakin' minor it begs the question: even if g as a corporate entity did not order this what is to prevent one of their biased engineers from adding that one line of code? (And they are biased, look at the political donations from g employees to the left)

[smallcompany]

After reading different opinions here, I thought about:

- impossibility of making factual conclusions when about Google
- how Google has evolved into a sort of amebae with no head nor tail you can catch it for
- lack of transparency and control of the private body that can make huge influence onto just anything
- how they can do (almost) anything they want, and pretend it wasn't for this but for that
- how everybody else can interpret the search result in the way desired or not

On the other side, if I for example take the third point, and bring it under a control, what good or bad does that do?

No matter about how much one screams about how fixed this was, and the other one about how wrongly it was interpreted, can any of us come up with more to prove anything?

Uh! My initial response to this was not like what I've just wrote above. Feeling like a sheep today...

[tangor]

Silicone Valley is all about H1-B visas (immigration). That is what keeps their expense down and wages contained. These tech companies know who will give them that .... and who won't. There is a very good possibility that position will be promoted, in this case by "doing nothing harmful to X" to protect those long term plans.

[ken_b]

When did politics become an acceptable topic at WW?

There's at least one mod and one admin in this thread

so it's clearly acceptable now and they should be able to answer my question.

[lawman]

Hello ken_b. This is a thread about Google algo relating to search. Granted the search involves politics. However, politics is collateral to the topic. Any conversation regarding whether or not G search on the topic is biased is valid.

[J_RaD]

Just more updates on more trickery

[theverge.com...]

[youtube.com...]

its a BUG.. you know, after they get caught.

and well....this

[i.sli.mg...]

money and power at play.


Its also interesting to find this thread is just kind of "hidden" if you are a public user not logged on.

HMMMMMMMMMMMMMMMMMMM *not shocked*

[JS_Harris]

Reverse the question then, if it annoys people. Is it possible for Google not to show biased results when the very same data collected by Google was used by CEO Eric Schmidt to launch a company with the stated goal of putting Hillary Clinton in the Whitehouse? Do many Google users(not webmasters) even know that Google has become political in nature from the top down?

We could ask Matt Cutts, but he works for the government's US Digital Service family now. I'm not sure he answers questions like he used to anymore.

Source of claim about Eric: [qz.com...]

[JS_Harris]

Today's Google news headlines involved the recent flooding in Louisiana. Google had headline links to six different stories from six different organizations. All six were anti-Trump along the lines of "Trump helps, for 57 seconds" and "Self serving as always". The articles themselves all praised Hillary's past work with "victims" of various disasters.

Then it struck me, Trump showed up in Louisiana, Obama was golfing and Hillary was sleeping/resting over the past few days per doctors orders. If those results weren't biased or politically motivated they'd have mentioned Hillary as a no-show, Obama's golf score and stories about what Trump said/did while he was there.

The bias is so strong in general that I think many people are being turned off by mainstream media, but it's also affecting rankings of "news" as well. This election won't be won by party worthiness, voter minds are being fought over in the media instead. I think the question of Google bias is a very good one to be asking, if not for this election then for the next, and the next. Google CEO's starting side companies to put specific people in the white house using Google data should also be examined. Times have changed.