Grammar Gripes: Studies Say … ?

A well-known Facebook group

The news was forwarded to me over email. “Grammar Police = Female Millennials.” And apparently 46 percent of American adults typically correct family or friends when they mispronounce words.

On August 20, Dictionary.com released the results of its online Grammar Gripes 2015 study (conducted by Harris Poll about three weeks earlier), and the press release got picked up by sites like PR Newswire, and then by the Associated Press and The New York Times. We here at Lingua Franca followed up, and the folks at Dictionary.com generously shared with us the full survey results.

As I read through the findings, I found myself thinking about what different kinds of studies can and can’t tell us about how people judge grammar and spelling “mistakes.” I have “mistakes” in quotes here because spelling is standardized enough that it is fairly straightforward to define what an error is; the situation with grammar is much more complicated. And some of what the survey calls “mistakes” are not: For example, they capture grammatical variation in the language (e.g., sneaked vs. snuck), and legitimate grammatical constructions people sometimes think are not grammatical (e.g., “I’m good”).

As a scholar, I am interested in people’s language attitudes and in how those attitudes may or may not have an impact on actual usage and language change. I have now spent years studying the nature of prescriptivism and arguing that linguists need to take prescriptivism more seriously as an object of study. As a result, I think there is real value in well-designed studies of language peeves and their effects. I also think we need to be very aware of how study design affects implications — and very careful about the scope of conclusions based on the data.

Here are some things the Dictionary.com survey tells us: what kinds of usage issues people think bother them when the bit of usage (e.g., commonly confused words like its vs. it’s, misspellings like definately) is presented to them in isolation; and how many people think that they often correct others, notice typos in menus and store signs, find misspellings and usage errors in social media and correspondence from friends, and are good spellers themselves.

You’ll see that I am emphasizing the verb think in that description because that is all we can know based on the survey’s design. We don’t actually know whether 46 percent of American adults typically correct family or friends when they mispronounce words; we know that 46 percent of them think or say that they do this.

The press release accurately summarizes that, according to this survey, “80 percent of American adults consider themselves good spellers,” and more women than men think highly of their spelling abilities. Whether or not these folks are good spellers is a different question. But here’s a conclusion from the press release that we can’t make: “Women notice the mistakes more, with three quarters (75 percent) saying they often find spelling errors in others’ writing, compared to 66 percent of men.” Given the survey’s design, we don’t know whether more women than men actually find spelling errors in others’ writing; we know that more women than men think that they often find spelling errors, or that more women are willing to say on a survey that they often find spelling errors. That finding is not uninteresting: For example, what language ideologies might encourage women to report that they are highly sensitive to standard language use? But we need to be careful about what the survey really says.

In terms of misspellings, the press release accurately represents one finding as follows: “When given a list of commonly misspelled standalone words, almost four in 10 Americans (38 percent) say they are most bothered by the misspelling of February.” It should be added that, from what I can tell, respondents could select as many misspellings as they wanted to from the list as ones that bothered them “most.” We also know that 49 percent of respondents say that they hear could care less used mistakenly for couldn’t care less (how many actually hear it is not known, at least from this study).

Another bullet point in the summary states: “The No. 1 commonly confused set of terms that is most bothersome to American adults in writing is their vs. they’re vs. there, coming in at nearly one in two (46 percent).” I would love to see here a caveat like the one above about “when given a list of commonly confused terms.” What’s the difference? It’s one thing to be presented with a list of common language peeves and then be asked which ones “bother you the most.” It’s another thing to see if people actually notice these kinds of usage issues in real life and are bothered by these usage issues.

Can we know that? Let me tell you about a really interesting study by my colleagues Robin Queen and Julie Boland that is being published this fall in Linguistics Vanguard. They wanted to figure out, among other things, if different types of errors in electronically mediated communication (in this case, email) elicited different kinds of judgments. In a carefully controlled study they presented undergraduate participants with emails responding to an ad for a housemate, and the students evaluated the writer on a social and academic scale. The emails might contain examples of one of three different types of errors: “typos” (e.g., mkae for make), “grammos” (traditional grammar “peeves”—e.g., mistaking there for their, could of for could have), and “hypos” (hypercorrections like between you and I, which at least sometimes signal increased formality).

Queen and Boland discovered that both typos and grammos led to more negative assessments of the writer over all, at about equal levels; in other words, contrary to their hypothesis, the participants did not see typos as less meaningful (i.e., just a slip of the finger of someone who “knows better”) than grammos. But when the researchers separated out the social evaluation (e.g., “This student seems similar to me”), only the grammos produced a significant negative effect.

The different salience of typos and grammos was confirmed by a second experiment that asked participants to correct a paragraph with typos, grammos, and hypos. Typos were corrected more frequently than either grammos or hypos. Participants noticed these more readily and/or were more confident about correcting them.

Strikingly, participants’ self-reported grammar attitudes did not show an effect on how they evaluated writers based on typos, grammos, or hypos.

It’s interesting to know what “errors” in spoken and written language people think really bother them when they are presented with a list of possibilities. It is even more interesting, I think, to know which kinds of “errors” people actually seem to notice in more real-life situations and how they judge writers as a result.

Language peeves have real implications for real people, and they merit continuing rigorous study. We still have more too (oops! grammo … be kind!) to learn.