Skip to content

A simple experiment in cognitive enhancement: what effect does the web and Google have on IQ?

by Michael Nielsen on November 8, 2008

A psychology experiment I’ve long wanted done is to give a group of people an IQ test where they have web access, and to compare the results to a control group without web access. I suspect those with web access would do quite a bit better, but it’d be interesting to see just how much better. “The web raises your IQ x points”, where x = … I’ve heard of a few experiments vaguely in this vein, but nothing seems to be quite like it. It could be run pretty cheaply on a University campus.

11 Comments
  1. I suspect much the same, but it may only show who gets more answers correct on an IQ test, not who actually has higher IQ (and IQ is certainly controversial in and of itself)… it’s sort of like asking whether people do better on open notes tests or closed notes tests. Then again, one could also argue that it’s not necessarily access to the web that has an impact, it’s how well you are able to find and use information from a sea of it, which is definitely an intelligence of sorts.

    There are always those who say the internet has made us dumber because we don’t have to “know” anything anymore – we can just google it. And those who say it’s not that we’re getting dumber, it’s just that the skills and knowledge that are valued are changing – less emphasis on hard facts and more on managing and integrating increasing amounts of information from many different sources.

    The study would certainly be easy to set up but it would have to be carefully framed.

  2. I should note that, on most traditional IQ tests, the speed that one answers is a primary contributing factor to the score. I think this is an absurd thing, anyway, but one would probably have to design new tests.

  3. Danielle – Good point. About 20 (!) years ago I saw a test that had unbounded time, so they definitely exist, but I don’t know how common they are.

  4. Shirley – Yes, the results would definitely be open to many different interpretations. Still, I think it’d be fun to just try it and see! It’s an instance where pretty much any outcome would be interesting.

  5. What makes this experiment not simple at all is to make sure, when selecting the test population and the control group, that all socioeconomic variables that correlate with internet use/lack thereof, and with IQ, are controlled for.

  6. Migeru, while that would certainly be valuable, albeit far harder, the base study I describe would still be very interesting. You might get, e.g., a larger effect with University undergrads than with people in their 60s; either result would be fascinating.

  7. The Vlad permalink

    Mike, if you don’t know the base IQ scores of the control and web-access groups, you cannot meaningfully interpret any differences between the groups’ post-experiment IQ scores, right? Is the putative difference due to one group being inherently smarter or due to your experimental manipulation?

  8. Vlad – That’s the reason to have a control.

  9. The Vlad permalink

    Yes, but you would need to establish, before the experiment, that both groups have equal IQ.

  10. Vlad: not with randomized selection of the two groups from the same population. That’s standard with these types of experiments, to combat exactly the effects you describe. Outliers might be a problem; I’d want to ask an experimental psychologist how those should be dealt with. With a large enough sample, I doubt it would matter, anyway.

  11. The Vlad permalink

    OK, I see what you’re saying. In practice, though, sample sizes in cognitive psychology experiments tend to be pretty modest (

Comments are closed.