Now my first question is, seeing as the log-rank and cox-R are 2 totally different methods, is Iwfal sure his analysis is universal generally to survival trials, and general to all kinds of analysis ?
1) Cox Regression is not a totally different method. It is, in essence, a particular type of Log-Rank regression.
2) My results are empirical, not theoretical. Therefore I cannot say with absolute certainty that they are universal. However I did run them with many different HRs, curve shapes, and trial sizes. The results were pretty consistent - in 1000 trials of the same size a Cox Regression has a median improvement in p value of greater than 2x. Sometimes it was less than 2. But it never approached a 0x improvement.
if the cox-R was universally favorable, then all the well-paid biostatisticians would know that as a fundamental, and would push for it every time.
a) I suspect that they do push and generally the FDA pushes back.
b) You have too much faith in the 'tech professional'. In my field I could name a Fortune 500 company that performed one of their most significant routine analyses in a way that violated one of Newton's laws. Generally only a 5% error - but since they were oblivious to it they could, and probably did, get burned on ocassions when it was bigger. And I could cite multiple PhDs who violated Newton's Law in different areas with equal ignorance - generally in such ways that anyone who grok'd freshman physics and spent 5 minutes thinking about it would know was wrong, but the PhDs just ended up as boiled frogs.
