This past spring, four of us here at GW Libraries had the privilege of attending the 2016 Code4Lib conference, featuring a wide variety of talks and discussions relevant to anyone interested in technology in libraries, archives, and museums.
The closing keynote was given by Gabriel Weinberg, the CEO and Founder of DuckDuckGo. If you're not familiar with DuckDuckGo, it's a search engine committed to not tracking you.
Tracking Your searches: Good and Bad
When you search using Google or other engines that track you, there's the obvious privacy issue around the company recording of all of your searches, but there's another aspect (let's refrain from judging it for the moment) which is that it affects the results of your search. Sometimes you may actually want that, but sometimes you don't. But let's first see when and why this happens.
You and I May Get Different Search Results
I'm going to use Google as an example, but this could apply to Bing, Yahoo, and other popular search engines as well.
Search engines that track you incorporate several factors into determining which results you see. If you're logged in to Google and haven't turned off the personalization settings, to the extent they can be turned off, then Google bases your results, and their rankings, on your previous searches (and possibly other information it knows about you from terms in your email, etc.) to try to present you with results it thinks you're likely to want and to click on. Other factors it takes into account include your location based on your IP address.
When you're hungry and want to quickly find something to eat nearby that you might like, you might want results that are localized and perhaps even take into account what it knows about your preferences. But when you're doing research for a paper, you may simply want the most objetive, consistent search results possible.
Here's an example: A Google search on "Obama" yielded slightly different results when I was not logged in to a Google account, versus when I was logged in to my (personal) Google account. The top news links were different: NBC, BBC, ABC, versus NBC, CBS, BBC; and a New York Times link was ranked considerably higher when not logged in, versus logged in:
One result of personalized results is the phenomenon referred to as the "filter bubble," a concept coined by Eli Pariser in his 2011 book. A filter bubble means that you're presented with results that tend to further reinforce your existing preferences, beliefs, and opinions. There is some controversy around the extent of the effects of this, but it has been a topic more in the forefront lately, particularly when it comes to social media and how platforms such as Facebook and Twitter determine which news items to prioritize in your feed.
Privacy, Tracking, Personalization and Other Search Engine "Features"
Let's check Wikipedia to get a rough sense of which search engines employ tracking, share information with third parties, and which don't:
From https://en.wikipedia.org/wiki/Comparison_of_web_search_engines#Digital_rights as of June 30, 2016:
Is the knowledge that your information might be shared with third parties, and that the search engine might be at least attempting to modify your browser settings ("browser hijacking"), worth the tradeoff of the benefits you derive from using those search engines? That's a personal choice, but it might be worth your while to try out a variety of search engines, paying attention to which track and which don't track.
Can't I just use Incognito Mode?
Incognito Mode seems to be somewhat misunderstood by many people. Incognito Mode is a browser feature that refrains from saving your browsing history and cookies in the browser itself, but if you're logged into Yahoo, Google, etc. within the incognito-mode window, they're still saving your searches on their side, and results may still incorporate your location and/or IP address.
Trackless Search Engines
One solution to concerns about privacy and objectivity is to consider using a search engine which doesn't track you. One of these is DuckDuckGo, which we mentioned earlier.
Libraries and Privacy
GW Libraries follow in the long-held library tradition of respecting and protecting patrons' privacy as well as providing objective search results when you use our research tools:
We won't share your circulation records, and records of electronic materials that you accessed.
We don't track you! When you search through the library web search interfaces, you will get the same results as anyone else in the GW Community, and the GW search engine is not tracking or saving anything about you. We wrote it, and the code that runs it is open source, so you can see it for yourself on github!
And last but not least, you won't get advertisements!
The only factor that can change your search results is whether you're using the GW Libraries search interface from an on- or off-campus IP address. This is because some of the resources, usually resources that GW pays to provide, are available to you as a member of the GW community, but not to the general public.
We do anonymously log search queries that come through the "All" tab (fondly known as the "Bento" search). The queries are anonymous; they are not associated with any user or even an IP address. We use these to better learn about our users are searching for - particularly the most popular searches - and we use what we learn to improve the research tools we provide.
Here's an example of a view that we as GW Libraries staff can see. Note that there's no information about who submitted each search: