[hmkpoker]

[ QUOTE ]
During the six months the law was enforced the number of admissions fell significantly (-16 admissions, 95% confidence interval -31.7 to -0.3), from an average of 40 admissions during the same months in the years before and after the law to a total of 24 admissions during the six months the law was in effect.

[/ QUOTE ]

The problem is that the -16 swing falls within the 95% confidence interval. If the 95% confidence interval is -31.7 to -0.3, that means it is 95% likely that a variance of that magnitude will occur, and -16 falls right in the middle. I don't see what's so significant.

If the study did reflect something meaningful, they would have established the boundaries of standard deviation (for example, they could state that swings of +/- X, around the mean (in this case, 40 MCIs/month), fall within two standard deviations, or ~95%). Something occurring outside that boundary would be worthy of attention. If the 95% confidence interval was 50 to 30, then drop occurring outside that range (like 24 MCI/month) would have a frequency of <2.5%

Even if this were the case, it's still far from establishing correlation, let alone causation. If you're going to correlate smoking bans with decreases in MCI incidents, you need a sample size greater than one. You need to compare the prevalence of changes in MCI incidents in towns that did employ smoking bans with similar changes in towns that did not employ smoking bans. This study does not reflect that. The variation could be attributed to anything.

For example, the increase of late nineties ecstacy use among kids happened concurrently with the popularity of the teletubbies, and declined similarly. However, it's clearly erroneous to say that Tinky-Winky was causing kids to take molly and dance all night.

Lastly, COME ON. Use a little common sense. If we are to believe that this evidence is indeed statistically valid, the conclusion of this study is that legally facilitated decreased exposure to second hand smoke in public places lowers the incidence of heart attacks by more than one third? AND that this is observable immediately?!?! Christ, it takes hardcore smokers years, even decades, before they start developing serious health problems, but somehow the absence of smoke in a bar results in an IMMEDIATE decrease of heart attacks?