Mike_A
Well-Known Member
I do not agree that the sample size is sufficient to cover all demographics. 1000 people is not a very large sample at all, and their confidence interval is +/-4%, which is to say in reality beer could be as high as 40% and wine as low as 31%. For their sample means to be so close (36% vs 35%) I would not settle for any less than a +/-1% interval. This would require a substantially larger sample size (couldn't tell you how large without knowing the standard deviation of the initial sample) but I would reckon it's in the neighborhood of 5,000-10,000 people.
Just talking statistics here, not facts. I find it funny that lower income and non-white participants showed preference for beer - I attribute this to the cost / quality ratio of cheap beer vs. cheap wine. I'd definitely say BMC products are still of higher overall quality than, say, Boone's Farm but that's kinda comparing apples to oranges.
Just talking statistics here, not facts. I find it funny that lower income and non-white participants showed preference for beer - I attribute this to the cost / quality ratio of cheap beer vs. cheap wine. I'd definitely say BMC products are still of higher overall quality than, say, Boone's Farm but that's kinda comparing apples to oranges.