5

I found a claim that literature search (in Google Scholar, ScienceDirect, Scopus, Web of Science, etc) should be performed in Incognito/Private Browsing mode.

According to the claim, search results may be influenced by browser cookies (presumably those about previous literature searches), and if a user account is used, they may be influenced by the scholar's information (interests, etc.) as recorded by the search engine.

If that is correct, authors of systematic reviews may need to report how they access literature databases and search engines.

Is there any evidence to support that claim?

  • And where did you hear this claim? Did you ask them for evidence? Who knows what Google is doing with their info, but, e.g., Web of Science is paid by my institution for use by the company - it is not in their (WoS) best interest to be messing with searches from our IP addresses since they would no longer be trusted (or paid). – Jon Custer Sep 24 '18 at 19:39
  • 2
    I added a source but it doesn't present evidence. – Orion Sep 24 '18 at 19:52
  • 3
    @JonCuster: Some people might consider a benefit if the search engine can learn from your past searches and direct you to content that is more likely to be of interest to you personally. In most cases, this would likely be more helpful than reproducibility of search results. I'm not sure why that should make the engine "not trusted". But it would mean that you'd need to take special steps to get a reproducible search when you specifically need one. – Nate Eldredge Sep 24 '18 at 19:57
  • 1
    All I saw was references to Google Scholar. If you have used even plain Google search, you would notice that, indeed, Google tries to tailor results for you based on past searches, figuring that they should indicate something about what you might be looking for. They are scary good in that way, so applying it to Google Scholar seems like a no-brainer. But, frankly, relying on Google Scholar to do serious literature searches just seems like a bad idea. Go talk to a real research librarian and do what they suggest... – Jon Custer Sep 24 '18 at 19:57
  • 1
    @knzhou Maybe that's a preference to set? imgur.com/a/YUgYv0E 116, 1373, 17532, 161, 963, 98. (I will note I get exactly the same results for that search incognito and not, however I haven't tested any terms closer to ones I would normally use) – Bryan Krause Sep 24 '18 at 20:10
2

It's probably good to resist treating any search engine as a "black box." That is, you probably should not simply accept the results without suspicion.

There are a number of factors that affect Google searches. Your search history, if you are doing the usual thing, is clearly one of them. Just a few pages explaining this.

https://www.bluecorona.com/blog/google-ranking-factors-2018

https://www.wsoaonline.com/does-your-internet-history-effect-google-search-results/

https://www.webpresencesolutions.net/7-reasons-google-search-results-vary-dramatically/

https://www.link-assistant.com/news/personalized-search.html

What it means is, private browsing would not hurt. And you probably should not stop with Google. You should probably use a bunch of search engines. Bing, Yahoo, DuckDuckGo, the archive searches on any journal archive you can get that is relevant, more if you can.

And you should be aware of the possibility that somebody has paid to have certain search results pushed up the list. Possibly on more than one search engine.

And then you should carefully examine the results to see if what you have achieved has validity. You should try to unravel any bias or detect any gaps as far as reasonable. Look at your data the way a hostile competitor would look at it and try to detect any flaws before you send it to a journal.

For example: Is there a point of view you are aware of that has been left out? Even if you think that POV is totally wrong, having the search engines decide for you could be embarrassing. Is one point of view clearly given undue precedence? Even if you prefer that POV you should be concerned if a search engine offered you only stuff you agree with. Is one source artificially cloned into multiple sources? Is a group of sources, that you know to be independent, artificially collapsed into a single source? And so on.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.