It may seem random, but it's heavily biased toward those who participate in politics. It leaves out people who have not contacted the senator. A better method would be to look at voter registration data, and send out surveys to random people on the list.