Comments on: Beer Drinkers More Irreligious http://brookstonbeerbulletin.com/beer-drinkers-more-irreligious/ Jay R. Brooks on Beer Tue, 22 May 2012 15:04:53 +0000 hourly 1 http://beerzine.com/?v=3.3.2 By: Bob Skilnik http://brookstonbeerbulletin.com/beer-drinkers-more-irreligious/comment-page-1/#comment-685 Bob Skilnik Fri, 23 Mar 2007 23:49:42 +0000 http://www.brookstonbeerbulletin.com/beer-drinkers-more-irreligious/#comment-685 Sam, I could get past the first sentence; "Alcohol and tobacco are more dangerous than illegal drugs like marijuana or ecstasy, according to a new British study." Ecstasy? C'mon, a proper pint is more dangerous? Sam,

I could get past the first sentence; “Alcohol and tobacco are more dangerous than illegal drugs like marijuana or ecstasy, according to a new British study.”

Ecstasy? C’mon, a proper pint is more dangerous?

]]> By: Sam http://brookstonbeerbulletin.com/beer-drinkers-more-irreligious/comment-page-1/#comment-684 Sam Fri, 23 Mar 2007 16:24:31 +0000 http://www.brookstonbeerbulletin.com/beer-drinkers-more-irreligious/#comment-684 Well the attack on alcohol (and by extentsion beer) is not over. Just wait till they see this (haven't looked at the actual study yet) http://edition.cnn.com/2007/HEALTH/03/23/drugs.report.ap/index.html Well the attack on alcohol (and by extentsion beer) is not over. Just wait till they see this (haven’t looked at the actual study yet)

http://edition.cnn.com/2007/HEALTH/03/23/drugs.report.ap/index.html

]]> By: Loren http://brookstonbeerbulletin.com/beer-drinkers-more-irreligious/comment-page-1/#comment-683 Loren Fri, 23 Mar 2007 11:17:34 +0000 http://www.brookstonbeerbulletin.com/beer-drinkers-more-irreligious/#comment-683 Hey Bob, E-mail the author this link. http://www.churchbrew.com/church_history.html Or even this one. http://brewmasterdave.blogspot.com/search/label/The%20Holy%20Trinity Gesundheit! Hey Bob,

E-mail the author this link.

http://www.churchbrew.com/church_history.html

Or even this one.

http://brewmasterdave.blogspot.com/search/label/The%20Holy%20Trinity

Gesundheit!

]]> By: Bob Skilnik http://brookstonbeerbulletin.com/beer-drinkers-more-irreligious/comment-page-1/#comment-682 Bob Skilnik Thu, 22 Mar 2007 14:08:57 +0000 http://www.brookstonbeerbulletin.com/beer-drinkers-more-irreligious/#comment-682 And then, Jay, there' this; “This isn’t just a brew pub, it’s a church,” http://www.sbcbaptistpress.org/BPnews.asp?ID=25221 And then, Jay, there’ this; “This isn’t just a brew pub, it’s a church,”

http://www.sbcbaptistpress.org/BPnews.asp?ID=25221

]]> By: Henry Halff http://brookstonbeerbulletin.com/beer-drinkers-more-irreligious/comment-page-1/#comment-681 Henry Halff Wed, 21 Mar 2007 23:59:05 +0000 http://www.brookstonbeerbulletin.com/beer-drinkers-more-irreligious/#comment-681 oh ... and one more problem: internal analyses. These guys were fishing through the data looking for effects. There's a good chance that they found some that arose simply by chance. oh … and one more problem: internal analyses. These guys were fishing through the data looking for effects. There’s a good chance that they found some that arose simply by chance.

]]> By: Henry Halff http://brookstonbeerbulletin.com/beer-drinkers-more-irreligious/comment-page-1/#comment-680 Henry Halff Wed, 21 Mar 2007 23:56:46 +0000 http://www.brookstonbeerbulletin.com/beer-drinkers-more-irreligious/#comment-680 You managed to peg many of the problems with this study. I can name one of your problems that is not, and a few that you missed. First, I don't think that sample size is a problem. The results were statistically significant, so evidently the experiment had the required statistical power except ... 1. Effect sizes were not mentioned. (They might have been mentioned in the article, but the journal's subscription system is not functional and you can't get a copy on-line.) We don't know how much of an increase in their measures that we are looking at. The effects, although statistically significant, might have been negligible. 2. All of the subjects were in the same program, which means that they very well could have interacted with each other. One requirement of statistics is independence of cases. Who knows, maybe one really charismatic patient got to a lot of the others in group sessions. 3. Regression to the mean. Subjects enter the rehab program when their drinking is at its most extreme point. Measure their consumption at any later point and it's probably going to be less, even if the program has no effect whatsoever. Other possible problems: flaky instruments (e.g., the Daily Spiritual Experiences scale) of unknown reliability and validity, all of the problems associated with self report, no controls, experimenter bias and influence. This study is a gold mine of bad methodology. Too bad I'm not teaching statistics any more; I could have gotten several exam questions out of it. You managed to peg many of the problems with this study. I can name one of your problems that is not, and a few that you missed. First, I don’t think that sample size is a problem. The results were statistically significant, so evidently the experiment had the required statistical power except …

1. Effect sizes were not mentioned. (They might have been mentioned in the article, but the journal’s subscription system is not functional and you can’t get a copy on-line.) We don’t know how much of an increase in their measures that we are looking at. The effects, although statistically significant, might have been negligible.

2. All of the subjects were in the same program, which means that they very well could have interacted with each other. One requirement of statistics is independence of cases. Who knows, maybe one really charismatic patient got to a lot of the others in group sessions.

3. Regression to the mean. Subjects enter the rehab program when their drinking is at its most extreme point. Measure their consumption at any later point and it’s probably going to be less, even if the program has no effect whatsoever.

Other possible problems: flaky instruments (e.g., the Daily Spiritual Experiences scale) of unknown reliability and validity, all of the problems associated with self report, no controls, experimenter bias and influence.

This study is a gold mine of bad methodology. Too bad I’m not teaching statistics any more; I could have gotten several exam questions out of it.

]]>