"If you are looking for very small differences, you need very large samples, but if you are looking for large differences much smaller samples are satisfactory."
That's true only if you're dealing one variable. Because colds can be caused by over 100 different serotypes of rhinoviruses, not all of which may be equally vulnerable to zinc treatments, you're really dealing with more than one variable here. I would be surprised if the clinical trials included any effort on the part of Dr. Hurt to 1) confirm that the subjects even had a rhinoviral infection and 2) if they did, which serotype they had. I would have been much more impressed with a clinical trial that used 100 people in each of several different geographic locations. That would have reduced the possibility that all were infected with the same rhinoviral serotype and is how any drug company would conduct such a trial (and before I hear the broken record response, yes, I know GUMM is not a drug company and Zicam is not a drug. I use this for comparison only.)
As for my knowledge of statistics, I am not a biostatistician or an expert in the field. However, I have spent the last 10+ years working in pharmaceutical research, so I think I have a better feel for how these things work than Dan "I'm an Electrical Engineer" Zimmerman. Oh, and the time I've spent working in pharmaceuticals was focused entirely on virology projects- ie HIV, HBV, HCV, HPV, Influenza AND (to a small extent) rhinovirus. |
