4

First edition, Prof Harrell's REGRESSION MODELING STRATEGIES, section 8.2, "How many parameters can be estimated?", page 150, quoting "If predicting survival time were of major interest, we could develop a reliable model if no more than about 354/15 = 24 parameters were examined in modeling."

QUESTION: Where does the "15" come from? Is there a reference in the survival analysis section of the text (Chapter 16) which describes?

Note: Apologies if this is explained in the second edition of the text. I have just ordered a copy.

Jan Galkowski
  • 166
  • 7

1 Answers1

4

It comes from earlier in the book. Section 4.4: Sample Size, Overfitting, and Limits on Number of Predictors (p. 72 in the 2nd Edition; the simulation study that came to this rule-of-thumb is discussed in the subsequent two pages.): enter image description here

Mark White
  • 8,712
  • 4
  • 23
  • 61