Recently my oldest childhood friend asked me an interesting question about Google search results – “When you type in a request on Google, it gives you a list of possible websites you might be looking for. These change as you continue to type. Are the prompts they list the result of algorithms that are user specific or general (the same for everybody)? Would you and I get the same suggested websites if we each typed in identical requests on our own laptops or is it based on our Google search history or even our entire browser history?”
I responded – “Both the prompts they list, and the actual search results are user specific. You are correct – Google results are impacted by their algorithms which read the search and browser history.”
Of course, I kept going – “Google is really not an information retrieval tool – it is a sales and marketing tool. Saving your history gives them more push for selling things via their advertiser/partner network. The ads you see are also based on this record of your searches.”
While there has been some criticism regarding the impact of personalization and user specific results, it remains a primary driver of what every web user sees, not just in Google, but across the web. When confined to online shopping this may not be such a bad thing, but when expanded to searches on virtually any subject, the danger emerges.
Early information retrieval engines and library-oriented online databases would give each user, anywhere in the world, the exact same result. The results were based strictly on the search terms and strategy – using Boolean, proximity and field-based (i.e., title) queries.
The earliest examples of personalization were really user-specific convenience tools. These included the ability to save searches or set-up alerts which would let you know when new records entered the database which matched your search strategy. Certainly useful and still objective.
The online environment began to change when personalization of search results was introduced and became widespread. Writing in “Search Engine Watch” in November of 2005, industry analyst Chris Sherman, said “Google’s personalized search reorders search results based on your history of past searches, giving more weight to topics that interest you.”
Personalization of search results means you are not seeing the full range of opinion. Results are similar to what you have used before. This is potentially critical as it occurs in every query and precludes seeing information that would broaden your view by showing the full scope of available content.
Research, at its heart, is designed to expand and add depth to your knowledge on any topic. Just reinforcing your existing point of view is counterproductive to this goal.
The original information intermediaries, librarians, have always known that different types of sources contain distinctive content. For example, news stories have a different emphasis than academic research. Rumors or gossip are far less authoritative than peer-reviewed journal articles.
The body of knowledge (BOK) for any subject allows for range of opinion on a topic, but it is identified in context. All BOKs encompass varying points of view and include things like theories, forecasts, and established facts. BOKs also identify bias as part of the factors that influence the presentation of the information.
It can be argued that virtually all published sources have some kind of bias. However, often the information from publications does not reveal their tendencies, leaving the reader in the dark.
To combat this libraries and journalism professionals have been publishing media bias evaluation tools. Many of these services are exceptionally useful.
Two examples of these assessment resources include the “Media Bias Chart” by ad fontes media and the “Media Bias/Fact Check” website.
ad fontes media “…rate[s] the news for bias and reliability using a rigorous methodology and a politically balanced team of analysts. Our focus is on analyzing the news content of articles and shows.”
The Media Bias/Fact Check homepage says – “We are the most comprehensive media bias resource on the internet. There are currently 3500+ media sources listed in our database and growing every day. Don’t be fooled by Fake News sources.”
These resources would be more effective if they were immediately available in a user-friendly way at the point of need, that is within the search results.
In a confusing time when information is amplified and accelerated by the connectivity of the internet, it is critical to use the tools that allow one to determine what is the best information on any topic. It would be particularly useful if popular internet sites would integrate these evaluation mechanisms into their feeds or search results. The ethics of being an information intermediary demand it.