How about just identifying via their IP and name, and also log their browser user-agent too, then when getting the results, query the database to find IPs which

How about just identifying via their IP and name, and also log their browser user-agent too, then when getting the results, query the database to find IPs which have taken part more than n times, and manually check for the signs that the entry was done by the same person or a bot? It's not perfect, but unless you limit each IP to one entry, people can game the system (creating new Facebook accounts, email accounts, etc.) anyway.