In other news, this is called the "biased sample" fallacy.

1. Sample S, which is biased, is taken from population P.

2. Conclusion C is drawn about Population P based on S.

http://www.nizkor.org/features/fallacies/biased-sample.html