When scientists at MIT and Northeastern University gave 94 MIT volunteers modified cellphones they were interested in determining, of all things, friendship patterns.
The Bluetooth-enabled phones logged the volunteers' calls and their proximity to other modified phones.
The study, reported by New Scientist, opens up new methods for social scientists and epidemiologists to determine how and with whom people interact.
The study, for example, determined that proximity to another phone during work hours during the work week meant little but that proximity for several hours on a Saturday night meant the two phone users were very likely good friends.
The researchers were able to predict to 95% accuracy the "friendship quotient" based simply on cellphone proximity.