This page is intended primarily for those wonderful folks who participated in the study I conducted in the Spring of 2007, but may also serve as a useful summary for the work for others.

What were you studying?

I was looking at how the perceptions that we form about other people are influenced by the way they pronounce words that end in -i-n-g.

So you weren't studying whether we picked the raffle or the charity at the end?

Nope. That choice was there just to make it appeal to more people.

You're sure? Because that would have been cool.

Sorry, that's really not what I was studying. But if you're interested, of the 173 people who finished the study, 50 chose the raffle and 123 chose the charitable donation. That meant that 5 iTunes gift certificates were sent out and $615 was donated to Doctors without Borders.

Does soliciting subjects through a social networking site like livejournal really work?

In this case, it worked great! I got the majority of those responses within a week of posting the study, and most of those were within the first few days. The downside, of course, is that it was kind of a one shot deal, since if I called on the network to distribute my study every time I had one, I wouldn't have many friends left. There are a few livejournal communities devoted to research studies, but I haven't tried them yet so I don't know what the results might be.

Ok, so what were you studying again?

I was looking at how the perceptions that we form about other people are influenced by the way they pronounce words that end in -i-n-g.

See, I'm a variationist, which means I study different ways people have of saying "the same thing". Which is a bad definition, since it's mostly not the same thing, but that's another story. Variationists, especially what are sometime called third wave variationists, are excited about social meaning, which is the idea that small differences in speech are tools we use to say who we are and what's going on in a situation.

In some earlier research, I showed that (ING) does carry social meaning, by showing that people evaluated speakers differently depending on which variant they used: -in or -ing. But how it changed their evaluations was different for who was talking, who was listening, even what was being said.

One thing that came up a lot was expectation. People said that they expected Southerners to say -in and they expected "non-accented" speakers (in this case, from California) to say -ing. This got me wondering about the idea of a default-- is there any way to figure out which member of a binary pair is meaningful?

What did you do?

If you participated in my study, you know that first I asked you a few questions about who you are and where you're from. Then I asked you to indicate how well a bunch of words described your mood at the time you were filling out the survey, like happy, peppy, tired, gloomy and so on.

After that, you heard 8 speakers, and answered some questions about them, like how educated or casual/formal they sounded, whether you would describe them as annoying or laidback, and so on. The speakers were all university students, recorded in interviews with me. Four of them (Ivan, Bonnie, Tricia and Robert) were from North Carolina and the other four (Sam, Jason, Elizabeth and Valerie) were from California.

Each of you heard only one example from each speaker, but there were 12 possible versions you might have heard. Each speaker had four different clips, saying completely different things. And for each clip, there were three versions:

-in: all the instances of (ING) were said as -in

-ing: all the instances of (ING) were said as -ing

none: all the instances of (NG) were covered by static bursts

The last thing you were asked to do, after hearing all the speakers, was to remember/guess which variant each speaker you heard had used.

Did we remember?

No! But really, who could expect you to? The question was more what things influenced what you chose. And as I thought might happen, the regional accent (Bonnie, Tricia and Robert were from the South) had a much bigger impact than the actual variant, which didn't matter at all:

Then what was the point?

This is the cool part! Looking at the social evaluations (intelligent, annoying, etc), I looked for cases where one of the (ING) variants (-in or -ing) had the same rating or percent response as the "none" variant while the other variant was different. These were the places where one variant was adding some meaning and the other wasn't.

The -ing form added meaning by making the speakers sound more intelligent/educated, more articulate and less hip/trendy and less like a student. It also had two interaction effects. First, in general, the better your mood, the more intelligent/educated you were likely to rate someone (and the grouchier you were, the worse-- as one person said to me, my talk should have been named "I'm grouchy, you're stupid"). However! When the speakers used the -ing variant they skipped out of this pattern, and were heard as more intelligent than the others because mood stopped mattering. The second interaction effect was neat, too: older listeners were less likely to call someone religious, but this was dampened when hearing the -ing variant.

The -in form added meaning to other aspects of social perception It made speakers sound more casual and less likely to be described as gay. It also had an interaction effect, where generally (for -ing and none) younger listeners were more slightly likely to call someone a jock, for -in versions, it was the other way around. My guess here is that this isn't about (ING) so much as different ages having different details of what a "jock" is like.

If you want the details, here are some slides with the numbers.

Wait, so you get paid for this?

This is actually pretty exciting stuff. The data you helped with showed that even though our conscious minds think of region as the big deal in understanding (ING), when we're just listening to people and forming opinions, those expectations are not so important (which is not to say region isn't impacting (ING) in other ways!).

The other thing is that in variation, we mostly talk about the meaning of a variable. People can (and have!) argued about whether (ING) "means" something like education and socioeconomic class or whether it means casualness/formality. Since both seem to be related to (ING), it makes sense to wonder if one is the "real" meaning and the other somehow piggybacking on it. These results suggest that maybe we've been thinking about it the wrong way. Maybe it's not variables (like (ING)) that have meaning at all, maybe it's variants (-in and -ing) and the meanings don't have to be the same for all the variants of a variable.

What are you doing with these results?

I just presented them last week at NWAV 36, and pretty soon I'll be writing them up into a journal article and sending it off, hopefully to be published. I'll add a link if/when that happens.