Many researchers have focused on vehicular ad hoc networks in recent years. Because of the nature of wireless networks, security of these networks may be compromised against attackers. This leads to losses such as loss of people's lives and loss of productivity in trades and payment services. One of the harmful attacks against sensor and ad hoc networks is Sybil attack, which attacker tries to forge some identities. These identities can be real ones belonging to other vehicles in the network or bogus identities made by the attackers. The purposes of creating dummy entities are for network disruption, traffic safety violation and, briefly, for the attacker's goals. In this paper, we describe a general classification of various methods for Sybil attack detection in vehicular ad hoc network. We then analyze advantages and disadvantages of the selected researches in each class to evaluate their performance with regard to safety and privacy issues. Copyright © 2014 John Wiley & Sons, Ltd.