Re: st: Midas- initial values not feasible=error

[Jannik Helweg-Larsen <Jannik.Helweg-Larsen@regionh.dk>]

(sorry, to repost, but :)

If someone with the newest version of Stata ( > version 12)  and/or a more powerful system could be able to verify the following with the user-written midas problem- it would be most helpful:

(ssc install midas)

with the following dataset from 32 diagnostic studies, I find that the midas program is able to proceed without problems on my system with the initial 23 observations, but fails  with 24 and more observations:

e.g:
keep in 1/23
midas tp fp fn tn, res(all)


SUMMARY DATA AND PERFORMANCE ESTIMATES
Number of studies =  23
 Reference-positive Units =  482
Reference-negative Units = 3310
Pretest Prob of Disease = 0.13
etc etc- output omitted

in contrast:
keep in 1/24
midas tp fp fn tn, res(all)

initial values not feasible
r(1400);

I found this blog on gllamm- could this be relevant?: http://faustusnotes.wordpress.com/2013/08/09/resolving-problems-with-multi-level-modeling-in-stata/

Here is the dataset again:

tp      fp      fn      tn
12      30      0       123
22      6       0       186
17      63      0       273
8       16      0       29
31      35      0       334
11      21      0       118
12      8       1       63
9       18      0       78
35      46      0       448
21      12      0       152
17      16      0       77
9       2       0       12
39      6       1       241
7       4       0       21
9       5       0       27
13      36      0       164
7       11      0       16
25      25      0       77
31      7       3       87
53      38      0       11
60      75      0       160
24      11      0       46
5       5       0       71
2       30      6       33
45      8       9       75
11      1       0       62
11      1       0       9
24      18      1       84
27      0       0       38
26      9       0       17
53      16      0       33
16      7       0       24

thanks


Jannik Helweg-Larsen, MD  DMSc
Senior Consultant
Department of Infectious Diseases and Rheumatology
---------------------------------------------------------------------------------------------------------------------
I have trouble with the user written program Midas, which performs
diagnostic meta-analysis

available at SSC: package midas ( Using Stata 11.2 MP, fully updated on
Win XP)

Using the dataset below I get the following error:

midas tp fp tn fn, res(all)

initial values not feasible







________________________________


Denne e-mail indeholder fortrolig information. Hvis du ikke er den rette modtager af denne e-mail eller hvis du modtager den ved en fejltagelse, beder vi dig venligst informere afsender om fejlen ved at bruge svarfunktionen. Samtidig bedes du slette e-mailen med det samme uden at videresende eller kopiere den.

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/