Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Steve Samuels <sjsamuels@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st:Survival analysis - dealing with Right truncated data |
Date | Fri, 17 Aug 2012 14:56:00 -0400 |
Dear Themba: I have no specific advice to offer because your options depend on what information can be obtained (or estimated or assumed) about employment at the firms from which your data set is drawn. Since you are at VU University Amsterdam, I suggest that you seek assistance from faculty there. A quick look at the University web site shows that sophisticated work is done in the Faculty of Social Sciences, at least. I'm sorry that I can't be of further help. Good luck! Steve On Aug 17, 2012, at 2:03 PM, Phakathi, T.R. wrote: Steve, is there a way out of this - what can be done? ________________________________________ Van: owner-statalist@hsphsun2.harvard.edu [owner-statalist@hsphsun2.harvard.edu] namens Steve Samuels [sjsamuels@gmail.com] Verzonden: vrijdag 17 augustus 2012 16:59 To: statalist@hsphsun2.harvard.edu Onderwerp: Re: st:Survival analysis - dealing with Right truncated data Because you utilize no information about those who did not fail, you can say _nothing_ about the impact of covariates on survival. Example: compare 2 groups 1. Your data Failures Group 1: 1 2 Group 2: 9 10 What can be said: of those who failed, failures in group 2 were later. But this does _not_ mean that survival was better in group 2. 2. Complete Data Group 1: Failures 1 2 Not Failed 11 12 13 Group 2: Failures 9 10 Not failed 0 Percent who failed through T = 10 Group 1 20% Group 2 100% Steve sjsamuels@gmail.com On Aug 17, 2012, at 10:15 AM, Phakathi, T.R. wrote: The dataset ONLY includes observations that have failed (Right truncation). For those that have failed (in this case all who lost employment), there are details on the risk onset and failure date including individual and firm characteristics. I would like to estimate the impact of the covariates on Survival. Are the commands distinct from “normal” survival data? If so what are the available commands (Non/Semi ¶metric)? May I tap into the wealth of your experiences Thank you Themba Phakathi * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/