Comments on: Death is different: More on event history models http://www.thejuliagroup.com/blog/?p=214 Words from the Prez Wed, 08 Sep 2010 08:10:27 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: Ian Cuthill http://www.thejuliagroup.com/blog/?p=214&cpage=1#comment-4312 Ian Cuthill Wed, 15 Jul 2009 18:48:21 +0000 http://www.thejuliagroup.com/blog/?p=214#comment-4312 Most providential encounter. Will add weasel sensitivity to Alzheimers and organ failure as potential hazards in my plan for eternal life. Most providential encounter. Will add weasel sensitivity to Alzheimers and organ failure as potential hazards in my plan for eternal life.

]]> By: Ashley http://www.thejuliagroup.com/blog/?p=214&cpage=1#comment-4307 Ashley Wed, 15 Jul 2009 15:30:24 +0000 http://www.thejuliagroup.com/blog/?p=214#comment-4307 Thanks for the chuckles from using the weasel attack example. As someone who took statistics a long time ago, I feel like I need more explanation how the two time periods in the weasel example get modeled (you don't define the terms in the Weibull model). Thanks for the chuckles from using the weasel attack example.

As someone who took statistics a long time ago,
I feel like I need more explanation how the two time periods in the weasel example get modeled (you don’t define the terms in the Weibull model).

]]> By: Mike http://www.thejuliagroup.com/blog/?p=214&cpage=1#comment-4305 Mike Wed, 15 Jul 2009 12:18:41 +0000 http://www.thejuliagroup.com/blog/?p=214#comment-4305 Please reconsider the second sentence regarding "symmetry thing around zero." One requirement of a parametric survival model is that the distribution fit, whether it be Weibull, lognormal, exponential, etc. While not symmetry around zero, determination of the appropriate parametric distribution follows the same notions as normality for a linear model. Please reconsider the second sentence regarding “symmetry thing around zero.” One requirement of a parametric survival model is that the distribution fit, whether it be Weibull, lognormal, exponential, etc. While not symmetry around zero, determination of the appropriate parametric distribution follows the same notions as normality for a linear model.

]]>