Those who seem honest aren't always.

Within a second (or less) of meeting someone, we're already making judgments about their personality - whether they're nice, smart, or even adventurous.

We're especially quick to discern whether that person is trustworthy. Unfortunately, we're usually wrong in that area.

Fascinating new research from the University of Southern California and Temple University harnesses the power of computer science to determine what behaviors make someone seem trustworthy and what behaviors indicate that someone really is trustworthy.

Spoiler alert: The two sets of behaviors don't match up.

For the study, researchers had about 300 participants pair up and role-play negotiations. All the while, the researchers were filming them and tracking their facial and body movements.

Specifically, participants pretended to be the owners of two antique stores and had to reach an agreement about six antique items. Experimenters told participants how valuable each item was.

At the end of the negotiation, each participant rated their partner on honesty. Experimenters also figured out how trustworthy each participant was, based on how many lies they had told, namely whether they had stated an untrue preference for one of the items.

Using the tracking technologies and software, the researchers figured out which behaviors trustworthy people displayed and which behaviors seemingly trustworthy people displayed. Here are some of the most interesting findings:

Overall positivity: Participants thought it indicated trustworthiness. It didn't.

Controlled smiles: Participants thought it indicated untrustworthiness. It didn't.

Joyous smiles: These were linked to objective untrustworthiness, but participants didn't catch on.

Contempt: Participants thought it indicated untrustworthiness. It didn't.

Head pose variability (i.e. how much you move your head): Participants thought it indicated trustworthiness. It was actually linked to untrustworthiness.

Word count (i.e. how much you talk): This was the only behavior linked to both perceived and objective trustworthiness. The more people talked, the less trustworthy they were and seemed.

To sum up: People are poor judges of trustworthiness, presumably because they misinterpret certain behaviors. Specifically, the authors say that people rely too much on facial expressions when trying to gauge trustworthiness.

These findings are important given that psychologists say first impressions are pretty sticky - though in some cases, they can be changed.

In other words, you might enter a negotiation with someone, assume they're being completely honest, and end up falling for a bunch of baloney. Or, you might turn down the opportunity to do business with a generally honest person because you mistakenly think they're trying to pull a fast one on you.

This preliminary research gives us some guidance as to what we should be looking for when we're trying to determine how trustworthy someone is. At the very least, it should encourage us to take the time to think hard about our very first impressions which, in all likelihood, are wrong.

BI