Casual Fridays: What does an emoticon do?

Last week we asked readers to rate a set of statements they might see on Twitter. The premise of our study was that sometimes it's difficult to decide whether someone is insulting you or complimenting you. But we were actually examining a slightly different question: what effect does an emoticon have on a statement? Can you make a negative comment seem "nice" just by adding a smiley or a wink afterwards?

Everyone saw the same 13 statements, presented in random order:

• That's the most incisive comment I've ever seen
• You are just so *interesting*
• Yawn
• You're as brilliant as you are attractive
• I agree
• That is teh roxxors
• A statement like that seems like it should be obvious. Why has no one figured that out before?
• Some people just don't understand what a wonderful person you are
• Oh, sure. People who share your political view are *always* right
• Wow. Just wow. That is the most amazing thing I have ever seen
• That's absolutely ridiculous. Only a complete idiot would agree with that statement
• Thank God you and your big mouth only have 140 characters to express your dull platitudes
• Please, tell me more

However, the emoticons each viewer saw were different. Three of the statements were followed by a smiley -- :-), three were followed by a wink -- ;-), three had an exclamation mark, and three had no extra punctuation at all. The last statement ("Please, tell me more") was followed by one of four less common emoticons: ;-P :-( :-/ :-0.

We had over 1,100 respondents, so over 250 people rated each statement with paired with each emoticon. Did the type of emoticon affect the perception of the intended meaning of the statement? Here are the results:

This shows the average ratings for all the statements when followed by each emoticon. The rating scale went from 1--very insulting to 5--very complimentary. As you can see, on average, each emoticon led to slightly higher ratings than the statements shown with no punctuation. Even an exclamation mark, on average, led to a statistically higher rating than the statements with no punctuation, and the smile was associated with the highest ratings of all.

But averaging all the ratings together might be oversimplifying. After all, an emoticon might mean different things in different contexts. Many people view the "wink" emoticon as an indication of sarcasm, suggesting the statement means the opposite of its literal meaning. This graph shows the effect of the emoticons when the statements are divided into complimentary and insulting groupings:

I called a statement "insulting" if its average rating with no emoticon was less than three, and "complimentary" if its average rating with no emoticon was more than three (The four "insulting" statements were "Oh, sure. People who share your political view are *always* right," "Thank God you and your big mouth only have 140 characters to express your dull platitudes," "That's absolutely ridiculous. Only a complete idiot would agree with that statement," and "Yawn.").

For the insulting statements, both the Smile and the Wink led to more complimentary ratings, while the Exclamation Mark's ratings weren't significantly different from statements with no punctuation. For complimentary statements, both the Smile and Exclamation Mark led to more complimentary ratings, while the Wink's ratings weren't significantly different from statements with no punctuation.

So adding a wink or a smile can enhance the positive perception of a negative statement, but a wink doesn't change the rating of a positive statement. Smiles and exclamation marks both improve positive statements.

But perhaps most importantly, simply adding a smile or a wink to an insulting statement doesn't actually make it complimentary. While the average ratings were higher, they were still significantly lower than the "neutral" 3, and significantly lower than the ratings for any of the positive statements, no matter what emoticon followed them.

One problem with these results, as many respondents point out, is that it's difficult to interpret the statements without context. But all respondents faced the same lack of context; clearly emoticons can't compensate for this lack of context. Perhaps the larger lesson is that if you're uncertain whether the person you're communicating with will understand your intention, an emoticon isn't going to help much.

Of course, there are many more emoticons than just smiles and winks, and the thirteenth statement ("Please, tell me more") offers us some insight into how they are taken:

There was no significant difference in the ratings for ;-P and :-(, but :-/ led to more negative ratings, and :-0 led to more positive ratings.

Our final question, thrown in for a bit of fun, turned out to be controversial. We asked for readers' "Favorite fictional communications device." As several people pointed out, a Tricorder isn't actually a communications device, and arguably smoke signals and drums aren't fictional. Undaunted, I present the results, for whatever sociological insight they might offer: