Here is Google’s video describing this new service:

While Google does not share its trade secrets, the ReadWriteWeb blog reports that, in part, this system relies on Freebase, “a structured database of semantic information. It maps synonyms to help Google understand the meaning of words.” It also incorporates content from Wikipedia, Google Books and the World CIA Factbook [1], which means it is seeking to be more authoritative. This has implications for online content providers. As the EverSparkInteractive blog notes:

Because Google is providing this information to users, you can bet your bottom dollar the search engine will expect the sites it top ranks to provide information like this. So, when you are writing content for your website, remember to include verifiable facts. [2]

While libraries specialize in providing authoritative content via free and subscription databases and other resources, will Google’s new approach recognize this and increase their visibility? Librarians wear many hats, and it can be quite easy to slap up web content without proper metadata, let alone add important links “above the fold.” The structure and organization of information on a library’s website — its information architecture — may or may not be carefully planned and executed. How will this affect its ranking by Google now?

Have you used the new KG? Have you found it helpful? Tell us!

Reference:

1.         Mitchell, J., Google Goes Back to What It Does Well: Finding Things. ReadWriteWeb, 2012. 2012(May 29): p. Online article or blog post about changs to Google’s search interface, and the addition of the Knowledge Graph, with Wikpedia-like content.

2.         Baumwell, A., Google Knowledge Graph: Will It Change the Face of Search? EverSparkInteractive, 2012. 2012(May 29): p. Online article or blog post about changs to Google’s search interface, and the addition of the Knowledge Graph, with Wikpedia-like content.

Google, Yahoo!, Bing, and Ask.com are by and large effective search engines for helping lay users get health and medical information. Nevertheless, the current ranking methods have some pitfalls and there is room for improvement to help users get more accurate and useful information. We suggest that search engine users explore multiple search engines to search different types of health information and medical knowledge for their own needs and get a professional consultation if necessary.

Researchers searched for “breast cancer” using the four top search engines, combined the top 200 non-redundant results from each source into a list, and then gave that list to eight volunteers to evaluate, scoring there relevance to research on a scale of 0-10 (low to high). Volunteers were all highly educated, most of them with backgrounds in science. They were provided with six “gold standard” sites for information on this topic to compare against before making their determination:

1. U.S. National Cancer Institute
2. American Cancer Society
3. Mayo Clinic
4. MedicineNet
5. Wikipedia
6. Susan G. Komen for the Cure

The volunteers tended to score the popular science and personal websites the highest overall, while scoring corporation and advertising websites the lowest. Researchers conclude that search engines do provide good information overall, though their sample is small, both in number of participants and topics searched.

This study has a number of serious flaws in it, and begs the question: is a little bit of mediocre research better than no research at all? First, the high education levels of the volunteer evaluators, plus their choices to score popular science websites above commercial websites indicate high levels of information literacy. Yet other research shows that over 90 million adults in the U.S. (36% of the population) have poor health literacy. How would people like this evaluate and select the right sources using a search engine? As any librarian watching patrons search will tell you, most people never look beyond the first three search results, let alone the first page of results. And search engines like Google employ algorithms to customize search results based on your previous searches, which may or may not give you the best resources.

Next, the researchers note the shortcomings of search engines in “ranking the websites according to their usefulness,” and recommend that “users apply multiple search engines when looking for medical and health information online, instead of using only a single search engine.” As a librarian and past computer consultant, I’ve learned that many people cannot distinguish between “the Internet” and a web browser, let alone utilize and critically evaluate search engines and the search results they generate. Librarians and researchers are the professionals best suited to handle that task.

Finally, while the “gold standard” sites probably contain all the basic information on breast cancer from the view of their medical expert, how readable and understandable is their content across different groups and ability levels? Had the researchers chosen to include a librarian on their team, she or he could have provided important insight and evaluation of these sites in this area. And why was a site like MedicineNet included while MedlinePlus was excluded from the short list? MedlinePlus is advertisement-free, offers information in Spanish and other languages, and is written at a level that can be understood by a wider audience.

While search engines will, generally speaking, help end users find helpful health and medical information, the will inevitably lead some people to inaccurate or misleading information. And with so many people having low information and health literacy skills, it is imperative that they be directed to quality, evidence-based resources for answers to their questions. This study does nothing to assist in that endeavor.

References:

Kutner, M. (2007). Literacy in everyday life: Results from the 2003 National Assessment of Adult Literacy.

Wang, L., Wang, J., Wang, M., Li, Y., Liang, Y., & Xu, D. (2012). Using Internet Search Engines to Obtain Medical Information: A Comparative Study. Journal of medical Internet research, 14(3).

73 percent of users said they would not be OK with an online search engine keeping track of their queries even if the data provides personalized results in the future. And 68 percent said they were not OK with targeted advertising because they don’t want their online activities tracked and analyzed.

Google is the preferred search engine for most online searching, though people surveyed don’t want information about their searches tracked. Yet unless you take steps to make sure this information is not collected, Google tracks it. And its search algorithms that include this information are what provides users with  the results they like. In a sense, we want it both ways: quality search results and privacy protections.

This is an area ripe for research: if we could define and control what information search engines and other online service providers can collect, would end users’ satisfaction with search engines change?

What do you think? Tell us!