The High Cost of Trust

Trusting information turned tragic when Wei Zexi, from Shaanxi province in China, died on April 12 after undergoing experimental medical treatment promoted on Baidu, the Chinese search equivalent of Google.

The typical hasty grab for digital information is seldom so terminal, but the common denominator is a failure to investigate.  What's different in Wei's case is that Baidu is criticized for failing to investigate the claims made by the advertisers it promoted in search results.

So, who's to blame? A search engine for failing to vet its advertisements? A student for believing information provided by a search engine?

It's always a problem of what or whom to trust. It cannot be assumed a search engine can be trusted, even if the ads at the top of the page are paid by a hospital. It should be interesting how the case will be resolved in China. Breaking Baidu's search engine monopoly, which has been suggested, isn't necessarily an answer. Training searchers to investigate the information for themselves is a more responsible and trustworthy solution.

Start investigating here

Information Fluency Investigative Challenges

Invisible Query

Time flies! I've neglected this blog for about 6 weeks.

Dennis O'Connor and I are deep into authoring a book on Teaching Information Fluency. Our deadline is the end of April.

Writing a book is a discovery activity for me. Last time I wrote this much was my dissertation and I discovered plenty about flow and mathematics while doing that.

This time, while it would seem I've traversed the topic of information fluency through this blog and the 21st Century Information Fluency Project website, there are still Aha! moments.

As I was thinking about the process of querying, it occurred to me that there's a lot more to it than translating a natural language question into a query. That's just the visible query--the one that search engine responds to. There's also an invisible query, the one you don't enter into the text box. The keywords or concepts that remain in your head.

These help you filter the results of the query. Some results are more relevant than others, not due to their ranking, but because you have some priorities in mind the search engine is unaware of.

It's generally ineffective to enter everything you're looking for in a search box.  It constrains the search and produces fewer results--sometimes none. It's better to submit two or more (keeping it a small number) keywords and scan the results for your invisible query.

Using one of our classic examples, "How many buffalo are there in North America today?", a good query is buffalo north america (bison is better than buffalo). Yet that's not really enough information to answer the question which is going to be 1) a number and 2) as recent as possible. That's the invisible part that you have to remember throughout the process. You choose results that satisfy 1 and 2; if not, you're probably not answering the question.

One premise of the Filter Bubble is that the machine is learning from us and will hone its output to our preferences. This becomes a harder task when we are not feeding the machine everything we have in mind. It may be a pretty good way to keep the Filter Bubble from encompassing us.

Think about what you're not querying that you are still looking for next time you search.