0

I get most of my news from the Google newsfeed, figuring it's a pretty balanced place because the feed is a collection of news sources and I know their respective biases; Google, I assume, doesn't inject bias into the algorithm (but it may queue off my inherent confirmation bias).

But there is one part on the sidebar that always gets me, I thought it would be different when I added a new Google account for some GCE work I had to do and accidentally logged into Google news, but it was exactly the same, the fact check sidebar always skews toward saying good things about left leaning politicians that are said by right leaning news sources, confusing satire sites or random viral social media posts.

I'm sure there are people who love to study this sort of thing, but is there a meta-analysis that we can do from this site that might show the percentage of claims one should legitimately be skeptical of coming from one viewpoint or the other?

8
  • I'm asking this as a social-conservative myself, who is super annoyed that the right-leaning-crap always bubbles up to the top, most of it is so stupid or pointless that it doesn't need to be fact checked (if it was, we'd have a lot more questions here). Sep 20 at 17:05
  • 5
    Questions on this site are only a small portion of the misleading claims out there and selected by the userbase.
    – Joe W
    Sep 20 at 17:36
  • @JoeW Do you think are they liable to skew in the same direction as the "professional fact checkers" it stands to reason that they're more likely to be things that are actually interesting, not "does Joe Biden have a button that turns him off?". Sep 20 at 17:55
  • Hard to say which is why I didn't post an answer.
    – Joe W
    Sep 20 at 18:21
  • 7
    I'd tend toward no, mostly because users post the questions. There's no reason to believe the question topics here are representative of anything other than our users' whims.
    – fredsbend Mod
    Sep 21 at 1:02
  • The Left and Right aren't really symmetric. They've got folks with differing backgrounds, personalities, behaviors, etc., leading to differing modes of expression and engagement. And it's even difficult to broadly generalize what those differences are, as both the Left and Right are composed of many different kinds of people. So while it'd seem neat for someone to do a robust, careful study about what such observables (like sentiments of questions from SE.Skeptics) might imply, it'd seem difficult to draw meaningful conclusions without some careful modeling and analysis.
    – Nat
    Sep 26 at 16:40
  • @fredsbend i.e. en.wikipedia.org/wiki/Selection_bias Frankly knowing the individual biases of a number of users here isn't that hard. Some post almost exclusively stuff that makes (just) one side of the US political spectrum look bad. But that doesn't even mean such users are proportionally representative of the US population etc.
    – Fizz
    Oct 6 at 19:09
  • @PeterTurner Your belief that google doesn't pick sides is unfounded. usatoday.com/story/opinion/2018/09/10/…
    – Ryan_L
    Oct 15 at 6:29
1

I think this would likely not make a good Skeptics.SE question.

  1. Predominantly, there is no notable claim here.

  2. If there were, it might deal with my next objection: the claim is so full of vague terms as to be untestable.

  3. You appear to be asking us to do the analysis, but no-one has any reason to trust the analysis of any of our users. Instead, we ask that answers link to analyses (which themselves should provide reasons to trust it - like references to support it's claims).

If you can tackle those issues, it could make an interesting (if controversial) question.

(Aside: I reject your argument Google doesn't inject bias into the algorithm, even if they don't deliberately inject bias into the algorithm. If they try to "maximise engagement" they may bias towards controversial articles. If they try to tailor to their users (or to the web-pages represented), they may bias towards articles of interest to their majority WEIRD users. Skeptics.SE has a notable bias towards questions of interest to US programmers, because of Stack Overflow, but I don't see any evidence of thumbs on the scale.)

3
  • 3
    yeah, I'm not thinking of asking it on the main site, just asking as a meta discussion. I think the "Programmer Bias" is pretty strong here! I used to think that Google was injecting bias via ML, they definitely tailor articles (guessing you don't see a lot of Pope news if you're not Catholic), but the "fact check side-bar" was identical between a new account and a 15 year old one. If the side-bar is identical for new and existing users, it's like the newspaper (possibly tailored to IP / region - but not tailored to users) Sep 23 at 14:55
  • Interesting. I do not read the Q as asking advice for any (possible) mainQ at all. More like a 'how can this site's posts be used outside of SE for 'gains''? // Google injects plenty of biases of all sorts in many directions. That's a given, at that level of abstraction. Sometimes even required by law to do so (de-listing/censoring certain results, EU-law 'right to forget' etc). Plus it's closed src, so who knows what exactly about it? Sep 24 at 17:32
  • "from this site" I read this as "users from this site". If that is meant to be "questions from this site", the answer is an easy no. Skeptics.SE is by no means a fair sampling of claims.
    – Oddthinking Mod
    Sep 25 at 9:27

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .