On top of already thinking that polling is generally useless for drawing any reliable conclusions other than “x people responded y way when asked z questions,” I feel like poll results like this and the growing support for marriage equality indicate that the earlier polls didn’t actually indicate disapproval with greater numbers but that the polled people don’t necessarily think a lot about the particular issues and they just determine whether they support them conceptually in a poll either arbitrarily or based on their perception of the most acceptable answer.
One data point may be meaningless, just as one summer does not make global warming. It’s the long-term trend that’s pretty clear there. One can argue a particular poll is skewed, but that can’t explain away momentum.
I don’t disagree that it seems like the attitudes of some are changing. I’m just saying that the polling doesn’t necessarily indicate what’s being concluded if people who don’t actually have an opinion on a topic are asked to opine on it or if people are substituting their perception of society’s dominant perspective (or that of their subculture or their family or their profession, etc…) for their own. It’s possible that the positive trend for some is going from nominal disapproval (via opinion substitution) to actual approval (or perhaps more opinion substitution). If the polling can’t adjust for these possibilities (or the numerous other ones I haven’t even thought of), I’m just saying that we should take all polls cum grano salis.
Marijuana should just be the beginning… I think all drugs should be legalized, and the current drug enforcement funds should be directed toward education about drugs and treatment of the addicted.
The war on drugs is dumb, and should stop altogether.
Wow. This chart should be like an intro chapter on how not to make charts. What is with the variable x-axis? We jump from 1979 to 2009, then jump another 2 years for the next data point, then a year, then 5 months, then 10 months, then 14 months.
I’m not debating if the data is right or not, but good god this chart is terrible. The slopes it shows are just outright lies.
Error bars? Curve fitting?
Um… no: not with the supplied data. Understand that I will bleed you with my statistical chops, but that this is merely a recombobulation of the above chart to eliminate the distortion on the x-axis. (1) No sample sizes, no error bars, and no confidence intervals either. (2) This is a descriptive, not inferential plot, so connect-the dots is reasonable, and given the small number of points to work with I am surely not going nonparametric regression/generalized additive model, so no curve fitting.
Now go sit in the corner and eat a cookie.
Well, at least that’s more tolerable. Though I feel like the little marijuana leaves did give it a certain something.
This topic was automatically closed after 5 days. New replies are no longer allowed.
