Why It’s So Hard To Detect Emotion In Emails And Texts

Short answer: we’re selfish. But there are things we can do about it.


Earlier this year, in a story on enigmatic email, the Wall Street Journal shared an awkward exchange between a consultant named Jill Campen and her boss, Marty Finkle. Campen sent a detailed email outlining a broad business strategy on a Thursday, only to get a one-word reply from Finkle the following Monday: “Noted.” Dismayed at the brevity, Campen replied again to ask Finkle if he was mad at her. It took a phone call to clear the air and establish that, far from upset, Finkle was pleased to clear the matter from his inbox so quickly and confidently.


We’ve all been Jill Campen and Marty Finkle at times: struggling to convey our emotions over email (or texts or tweets), and struggling to interpret the emotions of others. The difficulty of expressive writing isn’t new, of course, but what’s relatively recent is the overwhelming amount of electronic exchanges we have with people whose personalities we only know digitally. Without the benefit of vocal inflections or physical gestures, it can be tough to tell e-sarcastic from e-serious, or e-cold from e-formal, or e-busy from e-angry. Emoticons and exclamation points only do so much.

So we’re bound to make some wrong assumptions on both sides of the ether, and as behavioral scientists have found over the past few years, boy do we. The evidence has also given researchers a better sense of why we suffer so many digital communication breakdowns (short answer: we’re selfish) and what we can do about it (short answer: make some face or phone time).

Let’s start with message senders. A big problem people have when conveying digital emotions is often that they fail to appreciate there’s a problem at all. In one experiment from a 2005 study, test participants emailed 10 statements to a recipient. Some statements were serious, some sarcastic. These senders believed the recipient would correctly identify the intended emotion behind most of the messages. In fact, the recipients only identified seriousness or sarcasm 56% of the time, which isn’t much better than chance (below).

Via Kruger et al (2005), Journal of Personality and Social Psychology

Here’s the key twist: When the same messages were transmitted through a voice recording, the recipient interpreted the emotion correctly 73% of the time, just about what senders expected. Vocal tones captured the emotional nuance that email couldn’t. The researchers believe that when people type out a sarcastic line, they hear it in their heads as sarcastic, and thus fail to appreciate that others won’t hear it the same way. In other words, our overconfidence when it comes to conveying emotions in email “is born of egocentrism,” concluded the research team, led by Justin Kruger of New York University.

“If comprehending human communication consisted merely of translating sentences and syntax into thoughts and ideas, there would be no room for misunderstanding,” Kruger and company write. “But it does not, and so there is.”


Now let’s look at such misunderstandings through the lens of the message recipient. It’s well-documented among psychologists that when people lack information, they tend to rely on stereotypes to fill in the gaps. In the case of emails and other digital messages, the missing information tends to be a full appreciation of the sender’s personality. That’s why it’s usually clear when a friend or loved one is joking in a note or text, but not always clear that a remote colleague is doing the same.

Take another experiment from a different 2005 study (though also involving Kruger). This time, test participants rated the intelligence of a stranger based on answers to questions received via email. Since they never met the message sender, recipients had to rely on a (fictional) photo and brief bio. In some cases, the sender was a well-dressed Asian with a high GPA and a double major in physics and philosophy. In others, it was a white kid in a Metallica T-shirt with a middling GPA in hotel management who’d been a high school football player.

Shocker of the day (that’s sarcasm, just to clarify): test participants rated the Asian’s email answers as more intelligent than Metallica dude’s. But when the exact same answers were given to other test participants over the phone–with voice filling in some of the character gaps–the two strangers were rated as equally smart (below). So when we receive an email from someone we don’t know too well, we often revert to personality stereotypes, and in doing so raise the chances of emotional misinterpretation.

Via Epley & Kruger (2005), Journal of Experimental Social Psychology

Digital miscommunication wouldn’t be much of a problem if we always adopted the most optimistic or generous view of an ambiguous email or text. If we all took “noted” to mean “he noted that” instead of “he hates me,” we could all move on with our days. But that’s not what we do. Management scholar Kristin Byron of Syracuse University has written that misinterpretation tends to comes in two forms: neutral or negative. So we dull positive notes (largely because the lack of emotional cues makes us less engaged with the message), and we assume the worst in questionable ones.


This digital slide toward neutrality or negativity came through in a 2011 study led by psychologist Bradley Okdie of the University of Alabama. Test participants were paired up and instructed simply to converse and get to know each other. Some interacted face to face; others via instant message. The face-to-face interaction took more reported effort–you had to actually acknowledge and deal with another living being–but also resulted in more positive ratings of the partner’s character, and an overall more enjoyable experience.

The lesson is a little face or phone time can go a long way toward exchanging more personality information, forming more positive impressions, and reducing email awkwardness. Short of that, it can help to use concrete emotional words in an email (e.g. “I’m happy to say…”), or to clarify someone’s tone (“when you said that, I took it to mean…”), or if you must, to dispatch emoticons. Some companies have been known to include disclaimers saying that brief emails may give a “false impression of curtness or insensitivity”–though people misinterpret the disclaimers, too.

If nothing else, Byron writes, it’s at least important to recognize that “we are fallible as both email senders and receivers.” Noted.

About the author

Eric Jaffe is an editor at CityLab, where he writes about transportation, history, and behavioral science, among other topics, through the lens of urban life. He's also the author of The King's Best Highway (2010) and A Curious Madness (2014)