A sample size of 362... that WANTED to respond to the topic is quite a bit different than a sample size of 362 randomly selected participants.
Though, I do think most Locks probably would want a tank spec. I personally don't care either way.
I imagine everyone on the lock forums is a lock, and that if there were truly pro or against locktanking, they would voice their opinions.
Those are perfectly reasonable assumptions.
For fun I ran a small test:
H0: The proportion of warlocks that want tanking is NOT the majority, p0 = .5
HA: THe proportion of warlocks that want tanking is a majority, p(hat) > .5
1 proportoin z test, conditions are: sample size times proportion > 10 (129 is greater than 10), and that the size of the sample times one minus the proportion is great than ten (also 129), and that it is a simple random sample, i.e. that it is the warlock community that replied to this forum survey, which is safe to assume as the survey has been there for weeks and only warlocks would go on the warlock forum. There would be no cause of bias I can see as anyone who would be against warlock tanking would also express their opinions.
P value comes out as .000038
At even an alpha level of .001, the evidence is clear that a majority of WARLOCKS want warlock tanking.
I just used statistics at 2 AM for a game.
I hate myself a little more.