tag:blogger.com,1999:blog-8434369190746987531.post3627357251287348858..comments2019-05-11T15:14:02.214+10:00Comments on Grog's Gamut: Australia’s Unemployment Rate steady at 5.4%Greg Jerichohttp://www.blogger.com/profile/04956402439870441083noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-8434369190746987531.post-19177304045773862122012-11-08T17:41:28.465+11:002012-11-08T17:41:28.465+11:00Yes Austin that's why I noted that showing the...Yes Austin that's why I noted that showing the decimal places reduces the confidence that they are accurate. Not everything has to be scientifically rigorous, sometimes you display things just because it's kind of interesting. They key is not to treat it like it is on a level with the data that has greater rigor (which I don't.) <br /><br />And actually I find the trend edge effects interesting. <br />Greg Jerichohttps://www.blogger.com/profile/04956402439870441083noreply@blogger.comtag:blogger.com,1999:blog-8434369190746987531.post-6202165057431160222012-11-08T15:59:54.170+11:002012-11-08T15:59:54.170+11:00Two points:
a) Adding more decimal places doesn&#...Two points:<br /><br />a) Adding more decimal places doesn't add any extra information as the standard error is just too large. You'll find the standard error in the percentage is of the order of 0.1-0.2%. As is standard in every other science (except it seems the dismal one), you use one significant figure for your error and round your estimator to that order of magnitude. So 5.4% contains the useful numbers, the rest can be 99.9999% (or so) accounted for by noise.<br /><br />b) Edge effects in trend estimates are boring. Everyone knows about them and why they are there.Austinhttps://www.blogger.com/profile/01464734296901197429noreply@blogger.com