[R] Type III tests and Cox models

[Terry Therneau]

This last brought forth a good chuckle. Perhaps combine them into a single entry?

On a serious note, to have someone of significant experience say that they hadn't heard of 
SAS's method for doing "x", for nearly any x, marks a welcome climate change.  (And took 
nearly that long.)
The phreg procedure also added a "Scheffe test" of the overall hypothesis, by the way, 
which no one here has yet back-engineered to figure out what they are doing.  An F-test?

Terry T.

On 01/20/2014 04:18 PM, peter dalgaard wrote:
>
> On 20 Jan 2014, at 23:05 , Göran Broström <goran.brostrom at umu.se> wrote:
>
>> On 01/20/2014 07:02 PM, peter dalgaard wrote:
>>>
>>> On 20 Jan 2014, at 18:47 , Terry Therneau <therneau at mayo.edu> wrote:
>>>
>>>> The short summary: I was suspicious before, now I know for certain
>>>> that it is misguided, and the phreg implementation of the idea is
>>>> worse.
>>>
>>> A fortune candidate, if ever I saw one.
>>
>> OK, but please state clearly that 'phreg' refers to a SAS procedure, not the function with the same name in the  R  package 'eha' (I should have chosen another name, had I ever heard of SAS).
>
> That'll be the 2nd fortune candidate of this evening...
>
>>
>> Göran
>>
>>>
>>> -pd
>>>
>