tag:blogger.com,1999:blog-4840253472679669200.post8733867326807110639..comments2017-09-25T09:27:17.733+01:00Comments on Advertising Bollocks: Vicks - up next: the cure for cancerPJHhttp://www.blogger.com/profile/11331948749785269728noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-4840253472679669200.post-41037636030792416722010-11-10T22:09:42.565+00:002010-11-10T22:09:42.565+00:00"The same as the sample size on the cosmetics..."<i>The same as the sample size on the cosmetics adverts. </i>"<br /><br />I keep meaning to pick up on one of those, but there's so many to choose from with their rejuvenating properties of dihydrogen monoxide, and the ever-lasting effects of natrium chloride, that I've never gotten round to picking on the absurdity of the statistics they use.<br /><br />Not that it's beyond the wit of (say) Nivea's advertising department to bend statistics in <a href="http://www.asa.org.uk/Complaints-and-ASA-action/Adjudications/2010/11/Beiersdorf-UK-Ltd/TF_ADJ_49336.aspx" rel="nofollow">their own special way</a>:<br /><br /><i>"Nivea Visage Anti-Wrinkle Q10 Plus helps reduce the appearance of fine lines and wrinkles." [A large group of the woman's friends were then shown arriving and surprising her with a birthday cake.]<br /><br />"So it's no surprise that 37% of women feel more attractive now than they did ten years ago." [text on screen: 37% of 12,267 women interviewed]</i><br /><br />The 12,267 interviewed were not (interviewed as) customers using Nivea products. Sadly I couldn't find a video of the banned advert, otherwise that would have been another post for today.PJHhttps://www.blogger.com/profile/11331948749785269728noreply@blogger.comtag:blogger.com,1999:blog-4840253472679669200.post-53730629646705495912010-11-10T21:27:05.184+00:002010-11-10T21:27:05.184+00:00Usually you can extrapolate from the x% of y sampl...Usually you can extrapolate from the x% of y sample size to the whole population - but only when y is a reasonbly large value and does not contain a single or limited type of element. But when the sample size is 90 the error factor is so large as to make any extrapolation useless. The same as the sample size on the cosmetics adverts.SadButMadLadhttps://www.blogger.com/profile/17836368722377421009noreply@blogger.com