1. Pete Tar

    Pete Tar Moderator Staff Member

    Please give your personal tips on how to effectively use search to track down a relevant piece of inormation.
    What are the useful google search boolean terms (I dont even know if that's the right word? You know, the codes you use).
    What are the best picture sourcing methods?
    How do I find the original of something?
    If I find useful information relative to something but it is on a site that also has what I would consider bunk (eg, man-made global warming denial, poorly researched anti-vaccine claims), is it still 'safe' to use if it is the only source? (eg, Mick your debunking of the Libor/aurora quote was from a website that is AGW skeptic).

    How do I analyse the editing history of a wikipedia page? What is the best way to use wikipedia?
    how do I use the internet wayback machine?
    Any other resources I should know of?

    Thanks.
     
    • Like Like x 3
  2. Mick West

    Mick West Administrator Staff Member

    Last edited: Nov 20, 2013
    • Like Like x 5
  3. Mick West

    Mick West Administrator Staff Member

    I only use the - operator (i.e. the minus sign "-") to exclude terms, and the " (quotes) to group terms. The skill to finding things is to get the right combination to narrow things things down.

    I quote often will copy a long chunk, like ten words or more, and try it in quotes, like:

    "In politics, nothing happens by accident. If it happens, you can bet it was planned that way."
    https://www.google.com/search?q="In+politics%2C+nothing+happens+by+accident.+If+it+happens%2C+you+can+bet+it+was+planned+that+way."

    Then an example of excluding terms:
    "In politics, nothing happens by accident" -fdr -roosevelt
    https://www.google.com/search?q="In+politics%2C+nothing+happens+by+accident"+-fdr+-roosevelt

    I also often will restrict seraches to a site, or to a domain, with terms like:

    site:cdc.gov
    site:.gov
    site:.mil

    geoengineering site:.mil
    https://www.google.com/search?q=geoengineering+site:.mil

    For checking when something first came up I use Google Groups, and Google Books, and google news archive

    https://news.google.com/news/advanced_news_search?as_drrb=a

    Google Books is great for searching for the earliest data a quote was used, and also finding early instances of various topics, like persistent contrails.
     
    • Like Like x 2
    • Informative Informative x 1
  4. Mick West

    Mick West Administrator Staff Member

    It would be better if you can use something like Wikipedia, and then preferable several sites. It kind of depends on what is being said. If it's just pointing something out - like there being no actual evidence of a LIBOR connection - then it's reasonable to quote it, or you could just paraphrase it. I just did that post rather quickly - but the bottom line there is that there's no actual evidence of a connection, but there's been no "oficial denials", so there's not a lot of repute you can quote.

    But yeah, choose the most reputable source you can, and don't just repeat claims. I essentially verified the claim by not being able to find the evidence myself, but that's a little weak.
     
    • Like Like x 1
  5. Mick West

    Mick West Administrator Staff Member

    I don't do much beyond clicking on "history", and the "prev" link for each change will tell you how it differs from the previous link. I think there are better tools, but I've not used any.

    Wikipedia is fine for quoting on uncontroversial topics, like ice supersaturation. But it's best to see if you can quote the multiple original sources as well.
     
  6. Mick West

    Mick West Administrator Staff Member

    http://archive.org/

    Enter the url and then "Take me Back". You will get a calendar with blue dots for each time the page was stored (might not actually have changed), and a timeline at the top. Click on one, then you'll get the old page and navigation controls at the top. Sometimes a page is not available, try a different date.

    Usually the oldest is the most useful, but sometimes it's useful click through until you see a change. You can use the URL to link directly to an old version of a page.

    Be aware that the owner of a domain CAN remove all the old entries from archive.org, so if you find something interesting, make a copy.
     
    • Like Like x 1
  7. Mick West

    Mick West Administrator Staff Member

    Another useful tool is "WhoIs", which comes in many forms, but you can just use on the web to look up who owns a domain and when it was created:

    http://www.whois.com/whois/funvax.com

    Domain Name.......... funvax.com
    Creation Date........ 2011-04-03
    Registration Date.... 2011-04-03
    Expiry Date.......... 2013-04-03
    Organisation Name.... Ryan Harper

    Reverse IP lookups are also useful, to see what sites are on the same IP:

    http://www.yougetsignal.com/tools/web-sites-on-web-server/
    [​IMG]

    If you get a large number of results that means it's on a shared IP address, and is low budget. A few related sites means its a dedicated server.
     
    Last edited: Nov 20, 2013
    • Like Like x 3
  8. Pete Tar

    Pete Tar Moderator Staff Member

    great stuff thanks.

    I have a person I want to check up on. Gene Rosen, said to be a retired psychologist, sheltered kids at the Sandy Hook massacre.
    Some real ...POS... on ATS has said he finds him to be 'unconvincing' as an ex-psychologist, and is basically calling his integrity and version of events into question based on that.
    What do I use to search for proof on this person being what he says?
    Thanks.

    Edit..Tried google search gene+rosen+phd, excluded recent reports, but nothing...
     
  9. Mick West

    Mick West Administrator Staff Member

    I searched for doctor review which led me to healthgrades.com, searched for gene rosen conneticut

    http://www.healthgrades.com/provide...e=provider&loc=conneticut&locIsSolrCity=false

    Gave third result:
    Possible daughter? Unlikely to be that many Rosen name psychologists in Sandy Hook, and careers run in families.

    excluding enough terms:

    "gene rosen" "sandy hook" -shooting -hero -feeding

    http://genespetservice.com/pages/about.html
    also dating back to 2004, so he's been retired at least eight years.
    http://web.archive.org/web/20040309143227/http://www.genespetservice.com/pages/about.html
    Same thing in many business directories
     
    • Like Like x 1
  10. Pete Tar

    Pete Tar Moderator Staff Member

    Brilliant.

    (really, that was so satisying to be able to do. I'm actually a bit emotional. Thank you)


    edit.. I got post-banned by using the same words he did to question the credibility of the psychologist, to question his credibility as a human being...
    The facts made no difference to him anyway.
    Oh well.
    I need to stop going there.
     
    • Like Like x 1
  11. Mick West

    Mick West Administrator Staff Member

    More detail on my thought process: The filtering out was the key there. You want to find out about "gene rosen" in "sandy hook". So I start with:

    "gene rosen" "sandy hook"
    Which gives lots of stories about the shooting:
    [​IMG]

    So I quickly pick a word to filter, "shooting"
    [​IMG]

    You want to filter words that are as specific to the story as possible. I pick hero, no so much from the results, but just knowing about the story.
    [​IMG]

    So then I see there are lot of stories with the same phrase "just finished feeding his cats", so I filter based on the least common word there "feeding", although I could have done -"feeding his cats" if there were no uncommon words. This gives us several results about Gene's Trusty Pet Care, which I then verified was real by looking at the earliest version on archive.org.
     
    Last edited: Nov 20, 2013
    • Like Like x 3
  12. qed

    qed Senior Member

    With Wikipedia you can view the (state of) the page at any point in history, as well as the differences.

    On any page, click View History at the top of the page: you will see all the edits backwards in time (who made it etc.)

    On the left of each edit line is (cur | prev), with no cur for the most-recent edit and no prev for the original.

    Consider some edit (cur | prev) 14:44, 17 September 2013 ...

    • If you click cur you will see the differences between the 17 September page and today, after which today's page is shown. So the older the edit the longer the diff list. Note that the list show all differences from 17 Sept to today.
    • If you click prev you will see the differences between the 17 September page and the page just before that, after which the full page as on 17 September is shown. The diff list will always be smallish in this case.
    Registered editors are identified by user name while anonymous edits are identified by IP number. Motivations and arguments appear in the comments.
     
    • Like Like x 2
  13. Elfenlied

    Elfenlied Member

    Some more tools on wikipedia:
    on the left (depends on layout you use) of a page, you have

    Navigation
    Interaction
    Toolbox

    What links here: all articles linking to the page
    Page information: has some info that may be useful, nr of edits, nr of distinct authors and "Number of page watchers": high number of watchers may indicate it's popular, controversial, often vandalised, edit warring.. (watchlist on wiki is the same as watched threads here)

    At the bottom of "page information", you have

    External tools

    Revision history search: links to WikiBlame, a tool to search the history of a page for the editor who added a specific piece of text.
    Contributers and User Edits: link to toolserver.org, but seems to have problems at the moment..
    Page view statistics: 30, 60 or 90 day page view statistics, at any date in the past (up to dec 2007 it seems)

    Those pages/tools are also available via the edit history that qed explained in the previous post: on top of the page, you have:


    External tools: Revision history statistics · Revision history search · Contributors · User edits · Number of watchers · Page view statistics


    On the wikipedia tools page https://en.wikipedia.org/wiki/Wikipedia:Tools (can also be reached by typing wp:tools in the wikipedia search), you get quite a long list of tools, among those for page histories, there are a few that can be useful for analysing editor activity. When you think someone is pushing pseudoscience or CTs, or editors working in team...:

    Contributor edit counts and analysis:
    • userhist – user script, adds interface elements for isolating a user's changes to a page.
    • Contributors – lists edits, similar to page history, but can be sorted by contributor; easy to exclude different groups
    • User contribution search – finds all the edits by a user to a single page
    • Page history statistics by user:aka – builds an edit history overview page
    • Intersect Contribs – compares the edits of two editors in any Wikimedia Foundation wiki to see which articles overlap
    • Wikistalk – similar to Intersect Contribs, allows comparisons between more than two users and namespace selection, works only on the English Wikipedia
    • Editor Interaction Analyzer, shows the common pages that two editors have both edited, sorted by minimum time between edits by both users. Only works on the English Wikipedia.

    All these are really useful only for people quite active on wikipedia or in the politics of wikipedia, to find and repair other "damage" by a bad editor, or to gather evidence for a WP:ANI, get an editor banned, stuff like that..
    (btw: these are copied from the cestoda (tapeworm) wikipage, so the links relating to the page will go there, was easier than taking a picture :p)

    Google
    a few other google tips:
    find pages linking to a website or page, example:
    debunked link:metabunk.org
    (requires at least one searchterm)
    find pages/sites similar to one you know:
    related:en.wikipedia.org/wiki/Polysorbate gives other pages about polysorbate, related:en.wikipedia.org gives brittanica.com, howstuffworks, planetmath,... (related:meatbunk.org didn't give results..)
    search for numbers in a range:
    ..2000 up to 2000
    iphone $150..$400 any value between $150 and $400
    1500.. 1500 or higher
     
    • Like Like x 3
    • Winner Winner x 1
  14. Critical Thinker

    Critical Thinker Senior Member

    An interesting read with good points.


    http://lifehacker.com/how-to-find-evidence-to-support-any-argument-1454627815

     
  15. Astro

    Astro Member

    I want to share a technique I've used for locating the original sources of stolen astronomical images that have been re-purposed for spreading bunk. This is especially useful in cases where a google image search fails to find the original image, as is common with star field images.

    The first tool in this technique is Astrometry.net, which is a very powerful resource. It's at least as important for general astronomical debunking as any search engine. It will astrometrically solve any image that contains stars. The end result is that you will know what the coordinates of the image are, what the field of view is, and you can get the coordinates for any pixel within the image. Unlike most astrometric software that requires you to know the approximate field of view and/or coordinates of the image a priori, astrometry.net can solve an image automatically with no user input. There are limitations, of course; if an image contains very few stars, it will likely not be solvable. The online version (http://nova.astrometry.net/ ) limits you to 10 minutes of CPU time per solution, meaning if it hasn't found the solution in 10 minutes, it gives up.

    For wide field images a solution is usually found within seconds. For very narrow fields of view, or for very star dense fields, it can sometimes take more time. The site is also prone to fairly routine down time, which always seems to strike when you need to use it. For those reasons I've downloaded the source code from their website and installed it on my own home PC. The online version will save your results as a webpage you can share, and the "new-fits" file it generates can be opened in a fits viewer to view the coordinates of every single pixel in the image, as well as perform measurements of the angular size of objects in the image. I use SAOds9 as my fits viewer and you can download that here (I find that version 6 is more stable on my PC than version 7):
    http://ds9.si.edu/site/Download.html

    In some cases, performing the astrometry is enough for the purposes of the debunking; if an image is said to show a certain part of the sky or specific coordinates, often times you will find that stolen images do not even correspond to the claimed coordinates. Other times though it's even possible to use the astrometry to determine precisely when the image was taken, which can then rapidly lead to the discovery of the original source. Just about any time where you have an image that contains one or more real solar system objects, you can figure out when it was taken. Let me show you an example. Donny Gillson claimed this was an image of a "heavy mass object" (Nibiru):
    [​IMG]
    Cropping the lefthand side of the image you can easily solve it with astrometry.net. You find right away that the image is nowhere near "89 Leo" as claimed. But there's more; if you look closely at the left hand side of the image, you'll notice something there that is not in pre-existing sky surveys of the same coordinates, and I don't just mean the photoshopped "crescent planet" (whose phase angle is of course silly as well for being an image that was apparently taken under very dark skies). You can use the astrometrically solved image to download the Digitized Sky Survey (Palomar Sky Survey) image for the same coordinates using SAOds9. This will allow you to see if anything has moved or changed between the images, and in this case there is a comet there on the left hand side of the image. Using SAOds9 to get the comet's coordinates, you can now figure out exactly when the image was taken. You just need to figure out when a comet appeared at those exact coordinates. Precision is important here; although other comets may have come close to that region at various times, it's unlikely for lightning to "strike twice" on the exact same set of coordinates, especially if they're not on the ecliptic. You can also verify that the comet's morphology matches what you see in the image. This technique can also work when a planet or asteroid is in the image. Basically what you need is some solar system object with which to "date" the image. Planets (in wide field images containing stars) and comets work best of course, simply because they're more rare and unique.

    In the example we're looking at, I already knew from experience what comet it was likely to be based on the morphology and the location, Andromeda. Whether you have a likely suspect or not, you can confirm what it is by using a planetarium program like Cartes du Ciel or Starry Night. Just set the view for coordinates of the solar system object present in the hoax image, and set the program to rewind time. Eventually your "culprit" will pass through the coordinates. Using this technique I figured out that the above hoax image was originally taken on about September 18, 2010, give or take (according to the hoax it was taken in 2012). The comet was Hartley 2. Now, even if a reverse image search won't work, you can quickly find the real image with google. Just do a date ranged google search for comet Hartley images on or about September 18th give or take a day or two. Very quickly I found the real, original image, here:
    http://astroblogger.blogspot.com/2010/09/comet-103phartley-near-pgc-71451.html
    [​IMG]
    And as expected, it did not contain any "Nibiru," despite having been posted years before the hoaxed image. I've used this same technique a few times now, and it is very powerful when conditions permit its use (the presence of a solar system object in the image is necessary).

    Again though, even if the image you're dealing with does not have any real solar system objects within it, it can still be very handy. Sometimes an ordinary planetary nebula will be falsely labeled "Nibiru." You can quickly debunk that and correctly identify any such nebulae with astrometry.net. You can also use the resulting astrometric solution to download Palomar Sky Survey images of the same coordinates for comparison and show that the star or nebula in question has always been there.
     
    Last edited by a moderator: Feb 12, 2014
    • Useful Useful x 8
    • Informative Informative x 2
    • Like Like x 1
    • Disagree Disagree x 1
  16. Mick West

    Mick West Administrator Staff Member

    • Like Like x 1
    • Agree Agree x 1
  17. Jazzy

    Jazzy Closed Account

    Thanks for everything so far. :)
     
  18. Jazzy

    Jazzy Closed Account

    Something similar happened here:



    :)
     
    • Like Like x 2
  19. Pete Tar

    Pete Tar Moderator Staff Member

    I'm wierded-out by the 1 'disagree' to Astro's post - either a personal grudge, or a mis-clicked icon.
     
  20. Jazzy

    Jazzy Closed Account

    Me too.

    ( "I" before "e" except after "c" doesn't work for "weird". Weird. :) )
     
  21. JesseCuster

    JesseCuster Active Member

    If you use Google Chrome as your browser, this function is baked into the browser itself. Right click on an image and choose "Search Google for this image" and it'll open a new tab showing the same results as shown above. A great tool I have to say as it's amazing the amount of images you come across on fringe websites claiming to be something they're not.
     
    • Useful Useful x 3
    • Agree Agree x 1
  22. Critical Thinker

    Critical Thinker Senior Member

    A great article from David Dunning in the Conversation:

    Why the Internet isn’t making us smarter – and how to fight back

    Excerpts from the article...

     
    • Like Like x 3
    • Agree Agree x 1
  23. txt29

    txt29 Active Member

    Some tips for using Reverse Image Search:

    Google is usually the first try, but TinyEye, although lacking some of the features of Google RIS, is an excellent alternative and saved my day many times in the past. I noticed that Google RIS performs poorly for example when you have only a black & white, or a color-modified copy of the photo. Google will often fail to find anything, or will only show similarly colored images.

    I usually start searching in Google, when nothing is found, I try TinyEye. It often happened that TinyEye showed me a colored version of a B&W photo, or an edited version that Google did not see, but it still did not reveal the original, or the website I was hoping to find. When I then passed the alternative picture from TinyEye back again to Google, it suddenly offered plenty of hits, revealing the information I searched.

    I noticed also that the image-URL search does not work always well. It may be perhaps due to the limitation of the search engines bots on certain servers by the robot.txt file, or due to connectivity problems. If the URL search fails, do not give up immediately - it may help saving the image locally, and using the upload search function instead.

    This small trick with Google Image Search may perhaps help someone too:
    When in a country where the local version of Google (for example google.fr or google.de) overrides the .com version, I found the interface for "all sizes" of an image becomes very user unfriendly. The US version and also some other localized mutations (i.e. google.cz) show a bigger version of an image quickly on the same page via JavaScript, and you can navigate through the list with the arrow keys. But when you click a thumbnail on the unfriendly version (*.fr, *.de and possibly other countries), it opens the target page instead. It is very inconvenient and slowing down the search. I did not find any option in Google settings for it, but when it happens, I simply replace the local extension (i.e. "fr" or "de") in the URL by "us" ("com" does not work, because it automatically reloads the local page again).

    When both Google and TinyEye fail, you can try also KarmaDecay. It searches in the Reddit posts, but at the top it also lists links to other image searches - besides Google and TinyEye, also Bing, and then you can also inspect many properties of the file at ImgOps. On the ImgOps page, there are links for inspecting HTTP headers, for encryption analysis, for viewing geo tags and other EXIF, and much more - it is a useful tool for tracking the origins, or for revealing some hidden details about the author, server, location, or camera. ImgOps also lists two additional Reverse Image Search links - Yandex (Russian SE) and Baidu (Chinese SE), which can be quite useful too.

    Of course, the functionality of all search engines quickly changes, so what I write here for example about the poor performance of Google on B&W or modified images, may not be applicable in a few weeks or months. It is still useful to know that Google is not the only alternative, and that the other options sometimes bring even better results.

    Have a look, for example, at the reverse image search of Mick's avatar in Google an in Tiny Eye - the "all sizes" of Google show only completely unrelated faces instead of Mick, just at the very bottom it shows the 48x48 pixels icon. In contrary, TinyEye finds plenty of versions in different sizes, including one of 552x672 pixels at mickwest.com:

    ris-01.
     
    Last edited: May 10, 2016
    • Like Like x 2
    • Useful Useful x 1
  24. txt29

    txt29 Active Member

    An update about the Reverse Image Search - although Google's RIS gets a bit better with the time (for example it now finally recognizes Mick's avatar), unlike TinEye it still has troubles to find original colored photos if you have only a monochrome version of it. Its database also quite differs from the one of TinEye, so if you do not find an image with Google, TinEye.com is definitely worth the try.

    However, in the last time I started to use more and more the Reverse Image Search function of the Russian search engine Yandex (https://yandex.com/images/search). Not sure if it is valid for all kind of photos and images, but I already tried it with couple of hundreds of photos, and am getting valid hits with Yandex at the majority of pictures that Google fails to find. Also, except of well known objects or photos from main-stream media, I am typically getting more relevant results at Yandex than at Google, and also the function "visually similar images" seem to be working better at Yandex.

    So if you fail to find an image source with Google, there is a good chance Yandex or TinEye can still help you.

    Both TinEye and Yandex seem to handle better than Google not only B&W vs. colored versions of the same photo, but also photos that were more seriously modified (trimmed, merged with other photos into a collage, etc.).
     
    • Useful Useful x 4
  25. NoParty

    NoParty Senior Member

    This is a small thing, but something I use daily to streamline results served up by Google.

    On the far right of the google page is a "Tools" button.

    If you click it, you'll get a useful time option that is set to "Any Time" by default.

    This gives you a drop-down menu that will let you limit results to just the last hour
    (good for breaking stories"), day, week, month or year.
    It's quick and helpful if you're trying to avoid too much stuff or dubious redacted stuff.

    But it also gives a "Custom Range" option which is what I use the most:
    This morning I was looking for info on a legal case from 2004...and used Custom Range to
    only return results from that year. Results returned were not a flood, and were on point.

    Likewise, I was looking for info about how NPS fared in the new Trump budget proposal,
    but knew that yesterday's (4/3/17) announcement of Trump giving his first ($78,333.32)
    salary check to NPS would dominate results and create such a flood of info that it would
    take forever to sift through it for what I wanted. So I used Custom Range from
    3/1/17 to 4/2/17 on "Trump" "NPS" & "budget" and got my relatively few results without any
    pollution from yesterday's big headlines.
     
    • Like Like x 1
    • Agree Agree x 1
  26. Mick West

    Mick West Administrator Staff Member

    For a very quick filter to remove current news, I often just put in last year - you don't always need to bother with accurate "from" and "to" dates:

    For example:
    20170404-120558-sa1wx.
     
    • Agree Agree x 1
  27. Critical Thinker

    Critical Thinker Senior Member

    I have been fooled by fake images, so the article about Tips on How to Spot Fake Photos on the Web from Gizmodo provides a few extra suggestions on how not to be fooled. These tips likely have been covered in other threads on Metabunk, but figured that having them provided here would be appropriate.

    Excerpts:

    As mentioned in a prior reply... Reverse Image Search

     
  28. NoParty

    NoParty Senior Member

    Tiny :D point: I actually learned to use https://tineye.com (similar to 'Tiny') from Mick a while back...
    though maybe teaching for "TinyEye" would still get you to TinEye...
     
  29. NobleOne

    NobleOne Member

    Thank you guys for this thread. I would also like to hear how to do basic image editing fast and effective so I can use those techniques when posting here on metabunk.
     
  30. deirdre

    deirdre Moderator Staff Member

    there are a few threads in the How To forum (on main menu bar if you are on a computer). some recommended editing programs, how to do the comparison side by side photos etc.
     
    • Like Like x 1
  31. NobleOne

    NobleOne Member

    Thanks deirdre. I can see it. Mick had already done the job.
     
  32. Rory

    Rory Active Member

    How about finding specific quotes on there, like the Brzezinski one from the other day? How did you do that?
     
    Last edited: Oct 14, 2017
  33. deirdre

    deirdre Moderator Staff Member

    You don't. You google the quote and if you see a source link, you 'copy link' address then paste it into the wayback machine.
    pop.