Tuesday, September 18, 2012

Silos and Imagination


Sharon Wildwind

Scroll down to the bottom of the blog to see the search results that 11 people did the week after this blog was published. Click on the charts to see them in a larger size.

Suppose you and I research the same topic. We type the same search term into the same search engine. We get the same results, right?

Wrong.

Welcome to the age of personalized search results. Here’s the key phrase, “Every click is a clue.” Internet Service Providers (ISPs) can be traced, so as soon as we sign on, those data collecting people—see last week’s blog about mountains of data being collected and stored—know that I’m in Calgary and either you are, too, or you’re somewhere else. Exactly who we’ve chosen for our ISP tells them volumes.

Even the way we enter information personalizes our Internet use. Let’s say we’re buying the same item, from the same on-line site. We’re asked to enter our credit card number and, as it happens, we have the same kind of card, with the same spending limit, from the same bank. My credit card number is entered with two clicks because I copy-and-paste. Your number is entered with several clicks, pause, clicks, pause, clicks, pause, and clicks.

Best guess? I do so much Internet buying that I have my number on my computer to be copied-and-pasted. You do a lot fewer purchases and are typing in your number as you read it from your card. That conclusion—correct or not—might mean that we are offered the same product at different prices. Because the computer concludes I’m more likely to make future purchases, I get a lower price than you do so that I’m encouraged to return later and buy more.

What does that have to do with on-line research? Put millions of clicks together over time, feed that information into search engine algorithms, and the engine personalizes what search results we get. If I go to Wikipedia all the time, and you don’t, Wikipedia is likely to show up in my results as the #1 or #2 hit. It may show up on page 3 or 4 of your results. If you are a Facebook user, and I’m not, you’re more likely than I am to be referred to a Facebook page related to your topic.

It’s called being directed to an information silo. We go up and down the silo to comfortable sites, which we probably like a lot, but we don’t branch out.

What’s wrong with that? I like Wikipedia; I get Wikipedia. You like Facebook; you get Facebook. Sweet, eh? Non-challenging. No need to be exposed to information that might rock our world view.

Therein lies the issue.

Cross-fertilization feeds imagination. If I go looking for avocado recipes and migrate along the way to a YouTube video about the 1940 Tacoma Narrows Bridge collapse, I might get a poem out of the experience. On the other hand if I go looking for avocado recipes and the computer decides to keep me in the recipes and kitchenware silo, chances are that I’ll end up with a tasty meal, but no poem.

Don’t think it’s possible to migrate from avocados to collapsing bridges? You haven’t wasted nearly enough time surfing the Internet.

We can probably live with a limited view of avocado recipes. But what if, as writers, we need to research competing views about controversial topics, such as social justice, child abduction, or identity theft? If we can’t turn off the creation of information silos—and we can’t—what can we do instead?

Do multiple searches, for example, “social justice, negative” and “social justice, positive” should give vastly different results.

Use Internet contacts and social media sites to ask people from very different backgrounds to do quick searches for us and send us the results they get on the first 1-3 pages of hits. A white, middle-age woman living in a small Midwestern town might ask a twenty-something woman of color in New York city and a older man living in Australia to do the same search for her. And return the favor by doing searches for them.

Game the system. Don’t stick to the same links all the time. Go exploring. Ask people of different ages and backgrounds what sites they go to for news and information and occasionally visit some of those sites. Do searches for weird topics. Play with the system. The more complex a pattern we build with our clicks, the harder it will be to force us into silos.

After this blog ran, several people expressed an interest in testing it. The following week I set up a search "women in engineering" and gave people 6 days to do the search and send me the results. 11 people participated. No one got exactly the same results, though the #1 result was #1 for everyone, and the top 4 results showed up in the same order for a lot of people.

Demographic spread of 10 of the 11 people who did the same search.

11 people did 12 searches (one person did two) The only perfect score was the IEEE Organization was first on every search.

The first 5-6 searches were coded as to where they came in that order. Some people sent screen photos with more than six results, and those results from 6 to 10 are listed at the right side of the chart. The number in parenthesis is how many times they showed up in results.

Some people had lists of related searches, differentiated between sponsored (advertisements) and non-sponsored listings. Over half of the sites had a photo of women engineers coupled to a link to Wikipedia.
--------------------
Quote for the week

Growing up, I didn't have a lot of toys, and personal entertainment depended on individual ingenuity and imagination - think up a story and go live it for an afternoon. 
~Terry Brooks, American fantasy fiction writer

6 comments:

Sheila Connolly said...

So The System is telling us what it thinks we should know, or what it guesses we want to know? When does it plan to take over the universe?

Sandra Parshall said...

News flash, Sheila: The System is already in control of our lives.

Sharon Wildwind said...

Now, now, fellow bloggers. Let's have a little optimism this morning. There is always gaming The System.

JJM said...

Excellent observations, Sharon, and rather frightening to any researcher, including librarians. It might actually be worth experimenting just on one machine but using different search engines for different purposes for a while, then throw the same search through all of them and see what happens ... Thank you.--Mario R.

Sharon Wildwind said...

Mario, that's a great suggestion. I'd love to recruit a few people to conduct identical searches and see if we get different results.

JJM said...

Suggest a search, then, Sharon. Exact same words. I can run them through on the work 'puter, and here at home; I'm sure others here would be willing to do the same. Post the question as a status update on your Facebook page, get more examples. We can post the results here or FB. I suspect we might all learn a lot.--Mario