An old college friend and accomplished writer, John Scalzi, recently posted a list of writing tips for non-professionals, which I'd highly recommend for professionals and non-professionals alike. One of his most unusual suggestions is to "speak what you write" -- literally, to read your writing out loud before publishing, whether in a blog post or just an e-mail to friends. This, he argues, will not only help catch spelling and other errors (each of which Scalzi says decreases the writer's apparent IQ by 5 to 10 points), but also help you see whether you're conveying the meaning you intend.

So what does psychology research have to say about this notion? (No, not that typos decrease your IQ, but the larger idea that reading your words out loud will help you determine if your meaning is clear.)

A team led by Justin Kruger conducted a series of experiments on how we perceive each other's intentions in e-mail, and their findings do have some relevance to Scalzi's claims. One common problem in e-mails is deciding whether your correspondent is being serious or sarcastic. Taking Scalzi's example, most readers will realize that one of his observations was sarcastic: your IQ doesn't literally decrease when you make a spelling error. But what about the advice given by the aptly-named blogger Grumpy old Bookman, who in response to the much-hyped controversy over fabrications in James Frey's memoir, suggested that authors literally make everything up, taking no inspiration from the real world? Most commenters to that post clearly thought he was being serious, but I have little doubt that the post was intended to be sarcasm (I also think he anticipated that many readers wouldn't "get it" -- and that was part of the joke).

But do most writers actually accurately anticipate how readers will perceive the tone of their writing? Kruger's team tested sixty pairs of students at Cornell University, asking each person to choose 10 statements from a list of 20. Each person had a different list; on both lists, some of the statements were sarcastic, and some were serious. In separate rooms, one member of the pair typed each of the chosen statements into an e-mail message. The other member recorded the statements with a tape recorder. Each person guessed whether the message recipient would be able to identify the statements correctly as sarcastic or serious; then they listened or read their partners' messages and indicated whether they actually thought the message was sarcastic. Here are the results:

While both e-mailers and talkers thought most sentences would be read accurately, e-mail recipients couldn't judge whether sarcasm was intended -- their readers guessed their intentions at a rate no better than chance. By contrast, people speaking sarcastic messages were accurately able to guess when recipients would see the sarcasm. Message recipients were also asked to say how confident they were in their understanding of the message, and again, whether reading e-mails or listening to recordings, nearly everyone believed they had accurately judged the message's intent.

So talking appears to be a better way of conveying sarcasm than e-mail. But what about Scalzi's advice -- can saying what you write actually help you better understand how your written message will be taken?

In a new experiment, pairs of volunteers e-mailed each other as before, but before they guessed how their message would be taken, they recorded

the statements on a tape recorder. Half of the group read the statements as intended, using a sarcastic voice for the sarcastic statements, and a serious voice for serious statements. The other half read them using the opposite intonation: a sarcastic voice for serious statements and a serious voice for sarcastic statements. Here are the results:

When people read the statements with the same intonation as they intended to convey, they were wildly inaccurate at guessing whether readers would judge the statements' intentions correctly -- in fact, readers again were barely better than chance. But when e-mailers tried reading the statements with the opposite intonation, their guesses as to how readers would perform exactly matched actual performance. So here is a case where speaking what you write does appear to help you understand whether readers will read your message the way you intend it to be read.

Kruger and his team argue that their study demonstrates that writers are generally overconfident about what their readers will understand. While confidence about our writing matches our confidence about speaking, in reality, we're less able to convey those intentions in writing. Amazingly, readers, too, believe they can effectively judge the writer's intent, so the potential for miscommunicating in e-mail is amplified. The research also appears to support Scalzi's claim that "speaking what you write" can improve writing, with a caveat: to better understand the potential for misreading, you should try to read your words using an intonation opposite what you intend.

As for the link between spelling errors and IQ, more research will be needed before a definitive answer can be reached (seriously!).

Kruger, J., Epley, N., Parker, J., & Ng, Z. (2005). Egocentrism over e-mail: Can we communicate as well as we think? Journal of Personality and Social Psychology 89(6), 925-936.