Comments on: “C” for reform http://ukcia.org/wordpress/?p=84 news and comment about cannabis law reform from www.ukcia.org Tue, 07 Sep 2010 20:31:57 +0000 hourly 1 http://wordpress.org/?v=abc By: DFDuncan http://ukcia.org/wordpress/?p=84&cpage=1#comment-1486 DFDuncan Mon, 21 Sep 2009 23:07:29 +0000 http://ukcia.org/wordpress/?p=84#comment-1486 This is an excellent discussion and I enjoyed reading it. As a statistician, I would like to offer one bit of explanation about the statement you quote that, "'analysis during the development of the evaluation concluded that to be able to detect differences between the two samples would require a sample of at least 50 schools.'" You comment, "Perhaps that reads as if the effect of the drugs education program was not as great as they were expecting?" Actually, it refers to a power analysis, which is a standard component of the planning stage of any well-designed study. A power analysis is conducted in order to determine the sample size necessary to detect any actual effects that may be there to be found by your study. With too small a sample, you would be unable to detect a difference in average heights between pygmies and basketball players. With a very large sample any difference may be found significant even if it is trivial and coincidental. In this case a power analysis showed that the proposed sample of 50 schools was too small to detect any differences that were actually present. This would have been before they had any results. It was something they did right. But having done it, they should either have enlarged the sample or abandoned the intent to do the study. David F. Duncan, DrPH, FAAHB This is an excellent discussion and I enjoyed reading it. As a statistician, I would like to offer one bit of explanation about the statement you quote that, “‘analysis during the development of the evaluation concluded that to be able to detect differences between the two samples would require a sample of at least 50 schools.’” You comment, “Perhaps that reads as if the effect of the drugs education program was not as great as they were expecting?”

Actually, it refers to a power analysis, which is a standard component of the planning stage of any well-designed study. A power analysis is conducted in order to determine the sample size necessary to detect any actual effects that may be there to be found by your study. With too small a sample, you would be unable to detect a difference in average heights between pygmies and basketball players. With a very large sample any difference may be found significant even if it is trivial and coincidental. In this case a power analysis showed that the proposed sample of 50 schools was too small to detect any differences that were actually present. This would have been before they had any results. It was something they did right. But having done it, they should either have enlarged the sample or abandoned the intent to do the study.

David F. Duncan, DrPH, FAAHB

]]> By: UKCIA http://ukcia.org/wordpress/?p=84&cpage=1#comment-1482 UKCIA Mon, 21 Sep 2009 11:48:19 +0000 http://ukcia.org/wordpress/?p=84#comment-1482 Hi Andrew Yes, indeed it was £6million, not 60 - I've corrected the mistake. Thanks Derek Hi Andrew

Yes, indeed it was £6million, not 60 – I’ve corrected the mistake. Thanks

Derek

]]> By: Andrew Brown http://ukcia.org/wordpress/?p=84&cpage=1#comment-1481 Andrew Brown Mon, 21 Sep 2009 10:44:25 +0000 http://ukcia.org/wordpress/?p=84#comment-1481 Hi, I think the questions you raise about the design of the Blueprint programme are entirely reasonable, but just to be clear it wasn't a £60 million study (more like £6m). Would that we could persuade government to invest that much in prevention research, though I'd be pressing for <a href="http://preventionaction.org/node/1305" rel="nofollow">these standards</a> should that happy day arrive. Hi, I think the questions you raise about the design of the Blueprint programme are entirely reasonable, but just to be clear it wasn’t a £60 million study (more like £6m).

Would that we could persuade government to invest that much in prevention research, though I’d be pressing for these standards should that happy day arrive.

]]>