Dunning-Googler —

Could search engines be fostering some Dunning-Kruger?

Study shows that we think Internet searches are a sign we know more than we do.

Could search engines be fostering some Dunning-Kruger?
Getty Images | Aurich Lawson

Many of us make jokes about how we've outsourced part of our brain to electronic devices. But based on a new paper by the University of Texas at Austin's Adrian Ward, this is just a variation on something that has been happening throughout human history. No person could ever learn everything they need to know. But that's OK, according to Ward: "No one person needs to know everything—they simply need to know who knows it."

Over time, we've developed alternatives to finding the person who has the information we need, relying on things like books and other publications. The Internet simply provides electronic equivalents, right?

Not entirely, according to Ward's latest results. Based on data he generated, it seems that search engines now return information so quickly and seamlessly that we tend to think we remembered information that we actually looked up. And that may be giving us unjustified confidence in our ability to pull facts out of our brain.

The speed of search

Ward's hypothesis is based on the idea that we probably categorize the recall process based on how easy it is. Wading through all the extraneous information in a book to find the single nugget we require can be arduous, even when the book is right at hand. While it can sometimes be difficult to latch onto a fact in our memory, it's generally much more convenient. For the easy-to-recall items—like the lyrics to annoying pop songs from our high school years—it's often instantaneous.

Of the two, Ward argues, Internet searches are more like remembering something, in that they're generally quick, don't have a lot of extraneous information, and are displayed via interfaces that are easy to process. "Thinking with Google," he writes, "which delivers information as unobtrusively as possible, may simply feel more like thinking alone."

If that's the case, performing searches to get information may feel a lot more like successfully pulling something out of our memory. And that could be misleading, since successful searches would give us the sense that our memory is more expansive than it actually is.

To test this hypothesis, Ward created a variety of information-recall questions. He then had people answer them, either by memory or by using Google. Layered on top of this simple scheme were variations on the recall challenge that helped identify how people viewed a successful Internet search.

Our sense of self includes Google

The most basic experiment involved having people answer 10 questions using either their memory or an Internet search and then take a cognitive self-esteem test, which measures how the subjects felt about their mental capacity. Those who were able to use Google got more questions right. But they came away with an enhanced sense of their own abilities. They were also more likely than those who relied on memory to say that they would do well on a future test in which they couldn't use the Internet.

From there, Ward's experiments branched out. In this case, both groups were given the correct answers to the questions they faced, allowing them to judge their actual performance. They were then asked about their confidence in a future test, and again, Google users were more confident. In this instance, however, subjects were actually given that future test, at which point it became clear that the Google-inflated confidence was misplaced, since the people who were deprived of the Internet in the second round performed just as poorly as everyone else.

In another test, the people who relied on memory were told that they got eight of 10 answers right, regardless of their actual performance. The ones who believed this score came away with an inflated sense of confidence that was roughly equal in magnitude to the people who used Google. Yet another test showed that this self-confidence evaporated if the Google users were asked to write down any answers they could pull out of memory before using Google. In other words, if they were forced to reckon with their memory's limitations before using the search engine, the subjects didn't end up with an exaggerated sense of their performance.

Channel Ars Technica