Environmental Economics

From the inbox via the Limdep listserv at 10:40 pm:

I'm trying to estimate confidence intervals for WTP using K&R in WALD for a latent class model. There are 13 RHS variables and 2 classes. The LCM estimation runs fine, but I get an error in WALD ("wrong number of start values") when I try to run the K&R.
Any help is appreciated.
Thanks in advance.

My reply to the list (which I figured out after many hours of dim-witted trial and error) at 10:47 pm:

Try doing about ten or so at a time with multiple Wald commands. Seems like I [sic] there is a limit on the number of functions allowed per Wald.

Another reply to the list from, er, let's call him "Bill Greene" at 10:48 pm:

It looks like you need to change ;Labels=b to ;Labels=26_b.
That is assuming that there are 26 parameters in the estimated B.

Now, I'm feeling really stupid for not waiting to see if the correct answer would appear, as it usually does from the Limdep support staff, but at least I now know a more efficient way to solve the problem. But still, I'm wondering if I can deal with the global embarrassment and get some sleep.

Then at 11:04 pm:

Thank you both for the prompt reply.

It was 28 (13 coefficents x 2 classes, plus two class probabilities).
Changing ;Labels=b to ;Labels=28_b worked great.

John's suggestion was on target as well.
WALD seems happy with < 10 fns, but crashes on me for anything larger.

Whew, zzzzzzzzzzzzzzzzzzzz.

I'm also good at spending two or so hours figuring out the incorrect way of importing data into Limdep but maybe I won't share that trial and error with the Limdep listser (first, save your Excel data multiple ways [*.txt, *.xls] and import < X columns with each version until somehow it magically works; second, proceed to analyze your data without writing down exactly how you got the import to work; third, a week later, repeat the process).