Trusting the Internet

One of my readers (Hi, Dad!) pointed out to me that the chart correlating higher average state IQs with a tendency to vote for Kerry over Bush which I linked to in Saturday's entry is in fact false. I've corrected the entry accordingly.

I don't feel guilty about the link (though I should have done a Snopes check before doing so, at least), since the entry was mainly written to push my own opinion. You may agree or disagree with it (no one has, in my mailbox at least), I do recognize that having gathered few-if-any facts in support of my thesis, it doesn't hold up as a concrete argument. But then, if you're reading my journal for deeply-researched political theories, you have deeper problems than my providing unverified information via a link.

This does raise an important point about the Internet which many people overlook:

Verifying information found on the Internet is hard.

Wikipedia is an open-source encyclopedia of knowledge, and is widely admired by many on-line citizens. But how reliable is its information? Most of the authors are amateurs just like me, and there is no formal review process for articles. My impression is that people seem to consider articles credible if they seem well-researched, without considering whether they're substantially incomplete or just plain wrong. While there is useful information there, there's no way to know that the information is reliable. To my mind, this significantly undercuts Wikipedia's credibility.

Of course, should Wikipedia really be considered less credible than a print encyclopedia? Are the authors of a print encyclopedia more reputable because of the capital required to publish their articles? (I don't really have an answer to this question.)

I rely on Snopes.com to debunk Urban Legends that go around the net. But the same principle applies: Why is their information any more reliable than a piece of e-mail I receive? I assume that it is, and as far as I can tell Snopes has had a good track record. But I have no idea whether they're perfect; for all I know, a fifth of what I read there is factually incorrect.

I maintain the top-rated Google page about the fictional character Peter Wimsey. I provide both factual information and my own opinions, but I also provide a timeline of the stories which I've worked out, which is essentially a compilation of facts and implications. I regularly receive e-mails from people who disagree with me regarding this timeline. Sometimes I revise the timeline after doing further research and deciding that they're right. Other times I decide I'm right and leave things as-is. But I have a de facto high-profile page for people interested in this topic. How credible do people regard the page? What are my responsibilities towards proving my claims on that page?

I dimly recall that my friend Bill was involved in a project a few years ago to determine the reliability of Internet content. He probably knows more about these issues than I do (he's more of a techhead than I am). It's a challenging task, not least because no one wants to get bogged down in intense research or debate about the reliability of specific pieces of information. People want to learn facts and read interesting analysis and move on, for the most part. But how do you know really, that information you're handed is true?

Another problem with the Internet is that the Web provides authors with the ability to change their text without archiving a copy. While it's true that search engines often archive pages they crawl, it's unclear whether those copies get saved in perpetuity, and if text is changed very shortly after it's written - before it's crawled by any search engines - then it may not be archived at all. This provides the ability for authors to effectively retroactively change their words to make their arguments stronger, or different.

For instance, I changed Saturday's entry by replacing this text:

However, this site does illustrate a correlation between state IQ and a tendency to vote for Kerry rather than Bush. (It points out a few other interesting tidbits as well.)

I'm not the only one to think of this, as this site suggests (the main chart therein turns out to be a hoax, though).

Think about how much of this probably goes on every day. News sites break news prematurely, then yank articles after a few people have already seen them. Sites yank articles when someone in charge decides it doesn't fit with their agenda or business practices. eBay yanks auctions. That's just the big stuff. What about the paragraphs, the sentences, the names which get changed in articles which stay up? Much of this is innocuous correcting or copy editing, but what about the stuff that's not?

The net is important. But no one's really policing it.