Originally Posted by Oberon

The DPA survey couldn't possibly be statistically significant because of sample bias. Delta pilots who visit the FPA website are not a random selection of Delta pilots. I have no clue how the results would skew but it's undeniable the results would not be statistically significant.

The ALPA survey suffers from the same sample bias but it isn't really trying to be statistically significant, it to s trying to be democratic. In other words, your vote only counts if you participate. If DPA, DALPA or anyone else really wanted to have statistically significant view of Delta pilots at a particular point in time (views change with time) they would have to select pilots at random.

My statistics education ended a long time ago so I'm not sure what a significant sample size for a population of 12,000 would be. It would be much smaller then the turnout of either survey but would yield better results