Do people automatically mimic others, even strangers?

Yes, participants did naturally copy the confederate (who they'd only just met) as measured by face touching, foot waggling and smiling. Face touching only went up 20%, but rate of foot waggling went up by an impressive 50% when participants were inspired by another foot waggler.

Does mimicry increase liking?

In the second experiment Chartrand and Bargh wanted to see if all this foot waggling and face touching has any actual use, or whether it is just a by-product of social interactions.

The set-up: 78 participants were sent into a room to chat with a stranger (another experimental confederate) about a photograph. With some participants the confederate mimicked their body language, with others not. Afterwards participants were asked how much they liked the confederate and rated the smoothness of the interaction, both on a scale of 1 to 9.

Result: Mimicry did indeed work to increase liking. When their body language was copied, participants gave the confederate an average mark of 6.62 for liking (and 6.76 for smoothness). When they weren't being mimicked participants gave the confederate an average of 5.91 for liking (and 6.02 for smoothness). Not a huge difference you might say, but still a measurable effect for a change in behaviour so subtle most people didn't even notice it.
Do high-perspective-takers exhibit the chameleon effect more?

Since we're all different, some people will naturally engage in mimicry more than others. But what kinds of psychological dispositions might affect this? Chartrand and Bargh looked at perspective-taking: the degree to which people naturally take others' perspectives.

The set-up: Fifty-five students filled out a perspective-taking questionnaire, along with a measure of empathy, then they were sat opposite an experimental confederate, doing the same old face rubbing and food waggling routine from before.

Results: Participants who were high in perspective-taking increased their face-rubbing by about 30% and foot waggling by about 50% compared with the low-perspective-takers. Differences between people in empathic concern, however, had no effect on mimicry suggesting it was the cognitive component of perspective-taking that was important in encouraging mimicry rather than the emotional.