"Spammers often use hacked accounts to spread spam," explained researchers Ghita Mezzour and Kathleen Carley from Carnegie Mellon University.

Modelling the behaviour of hacked accounts, the team found that spam distribution dynamics were different from what was predicted.

Hacked accounts were more aggressive in sending spam generated automatically by malware than individually propagated spam. This aggressive behaviour of hacked accounts caused spam to reach more people faster.

Repeated and forcibly sent spam to one's inbox was the online equivalent of word-of-mouth and was propagated further.

“Spam coming from a known email address was often more believable than that coming from an anonymous or scandalous source. Understanding and modelling the effect of the behaviour of these accounts is important to reducing spam and attacks on social networking sites," said Mezzour.

The research was published in the International Journal of Security and Networks.