convenient, flexible and yet entirely impirical
January 28, 2022 8:56 AM Subscribe
Statistician David Cox, known among other things for the proportional hazards model, died last week [paywalled WSJ].
Indeed. Sadly, I was not that clever. Just bad at spelling. Thanks.
posted by eotvos at 9:25 AM on January 28, 2022 [1 favorite]
posted by eotvos at 9:25 AM on January 28, 2022 [1 favorite]
. I used that model for an analysis with leeches once upon a long time ago...
posted by aeshnid at 11:10 AM on January 28, 2022 [1 favorite]
posted by aeshnid at 11:10 AM on January 28, 2022 [1 favorite]
I dropped out of a PhD program in "biomathematics" partly because their idea of biomathematics was nothing but biostatistics, and the "bio" was almost nothing beyond patient outcomes.
And I would say the linked "proportional hazards model" paper is almost a platonic ideal of a clear, well-motivated, and forceful presentation in ordinary language as far as is humanly possible, of an important idea with far-reaching consequences which are felt in the daily lives of almost all people in the developed world — and I got an inkling of a hint of a glimpse of an understanding of that idea. At best.
I like to think people can grasp the things that end up shaping their lives and make good decisions about them, but ... it’s hard to preserve that optimism sometimes.
posted by jamjam at 11:43 AM on January 28, 2022 [4 favorites]
And I would say the linked "proportional hazards model" paper is almost a platonic ideal of a clear, well-motivated, and forceful presentation in ordinary language as far as is humanly possible, of an important idea with far-reaching consequences which are felt in the daily lives of almost all people in the developed world — and I got an inkling of a hint of a glimpse of an understanding of that idea. At best.
I like to think people can grasp the things that end up shaping their lives and make good decisions about them, but ... it’s hard to preserve that optimism sometimes.
posted by jamjam at 11:43 AM on January 28, 2022 [4 favorites]
Two of the last 3 years of my life was devoted to work with a Cox proportional hazards model - I had no idea Cox was still alive, and was surprised to see him pop up here on Metafilter.
posted by peacheater at 12:40 PM on January 29, 2022 [1 favorite]
posted by peacheater at 12:40 PM on January 29, 2022 [1 favorite]
📈 .
The ability to answer a question like "how long until _____?" instead of arbitrarily picking a timeframe and asking "does _____ happen in this arbitrary time frame?" is huge in many different fields, and this class of models is either the grandaddy of survival methods, or maybe the cool uncle of them.
posted by adekllny at 4:59 PM on January 29, 2022
The ability to answer a question like "how long until _____?" instead of arbitrarily picking a timeframe and asking "does _____ happen in this arbitrary time frame?" is huge in many different fields, and this class of models is either the grandaddy of survival methods, or maybe the cool uncle of them.
posted by adekllny at 4:59 PM on January 29, 2022
I've used that model with psych data, political data, and marketing data.
.
posted by lathrop at 6:00 PM on January 29, 2022
.
posted by lathrop at 6:00 PM on January 29, 2022
« Older Inside the Cathedral of a Language | #ThresholdDay Newer »
This thread has been archived and is closed to new comments
posted by PresidentOfDinosaurs at 9:19 AM on January 28, 2022 [2 favorites]