Re: st: RE: question

 From Scott Cunningham To statalist@hsphsun2.harvard.edu Subject Re: st: RE: question Date Thu, 3 Nov 2005 06:47:26 -0500

Thank you - that is the relevant FAQ for my question. Sorry about that title, too. I didn't even notice that I'd put that until after it showed up on the listserv.

On Nov 3, 2005, at 6:38 AM, Nick Cox wrote:

First, note that the title "question" is not informative.

You have two questions. On the second, there is an FAQ

FAQ . . . . . . . . . . . . . . . . . . . . . . . Replacing missing values
2/03 How can I replace missing values with previous or
following nonmissing values?
http://www.stata.com/support/faqs/data/missing.html

. search missing

points to this.

On the first, I am no survival expert,
but absent a view of all your data your question looks very
difficult to answer. But you can at a minimum compare the results on
non-missing data and on data including imputations to see
what difference it makes.

Nick
n.j.cox@durham.ac.uk

Scott Cunningham

```I am estimating a hazard, but I am missing data for certain time-
variant covariates when t is very small.  Seeing data will help me
explain this.

+--------------------------------+
|       sr   _st   _d   _t   _t0 |
|--------------------------------|
1. |        .     1    0    1     0 |
2. |        .     1    0    2     1 |
3. |        .     1    0    3     2 |
4. |        .     1    0    4     3 |
5. |        .     1    0    5     4 |
|--------------------------------|
6. |        .     1    0    6     5 |
7. |        .     1    0    7     6 |
8. |        .     1    0    8     7 |
9. |        .     1    0    9     8 |
10. | .9738716     1    0   10     9 |
|--------------------------------|
11. | 1.081043     1    0   11    10 |
12. |  1.21261     1    0   12    11 |
13. | 1.125889     1    1   13    12 |
14. |        .     1    0    1     0 |
15. |        .     1    0    2     1 |
|--------------------------------|
16. |        .     1    0    3     2 |
17. |        .     1    0    4     3 |
18. |        .     1    0    5     4 |
19. |        .     1    0    6     5 |
20. |        .     1    0    7     6 |
|--------------------------------|
21. |        .     1    0    8     7 |
22. |        .     1    0    9     8 |
23. | .9738716     1    0   10     9 |
24. | 1.081043     1    0   11    10 |
25. |  1.21261     1    0   12    11 |
|--------------------------------|
26. | 1.125889     1    0   13    12 |
27. | 1.080463     1    0   14    13 |
28. | .9527455     1    1   15    14 |

Notice that for all values of t=0 to t=9, I am missing data for sr.
My first question is when I -streg-, what effect does this
have on my
estimation?  None of my respondents complete their respective spells
prior to t=10.  Secondly, assume I wanted to fill in those missing
values with the same value as t=10.  How would I do this?

```
```*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

```
```*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```