0
$\begingroup$

Given data with a binary outcome, i.e.: $0$ = no event, $1$ = event

which can be modeled with logistic regression, how then do we understand the following logic:

  1. data-1 (fully observed) $->$ estimate of maximum-likelihood coefficients $\hat{B}$
  2. data-2 - no outcome $\hat{B}_{data-1}$ $->$ estimate probability of outcome
  3. data-3 = data 1 & data 2 $->$ update $\hat{B}$ in a new logistic model.

What does this mean? Particularly when the dependent variable is changed from binary $(0, 1)$ to a continuous probability i.e. $(0 < P <1)$.

This logic follows that in a paper I have been reading.

Is this possible to reproduce these steps in SAS?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.