Hi,

Can anyone show me how to do the logit model overl excel? Here is the
sample data. Many many thanks!!

Flu Vaccine=1 EDUC AGE FEMALE Not Married
0 10 43 0 0
0 12 38 0 0
1 12 58 0 0
1 14 42 0 0
1 8 56 0 0
1 17 41 0 1
1 17 26 0 0
1 14 24 0 1
1 14 34 0 0
0 12 53 0 0
0 14 20 0 1
1 16 29 0 1
1 16 48 1 0
0 13 22 1 0
0 13 35 0 1
1 12 43 1 0
0 16 39 1 0
1 14 42 0 0
1 14 41 0 0

I'm reluctant to answer this, even though the answer is at the tip of my
fingers.

1) Do you know what a logit model is? Much less, how to use it? If
*that* is the problem, grab a decent book on econometrics, say (such as
that by William H. Greene), or on marketing research. Or, of course,
immerse yourself in a web-search; keywords "logit model estimation" seem
to do the trick, yielding plenty of stuff for anyone with a solid
high-school education in math that they still remember <g.

2) You can use Solver to great advantage here, if you know how to set up
your objective; a log-likelihood, I assume, is appropriate for
underlying assumptions you are likely to be willing to buy into. But
this takes me back to (1): do you really understand WTH you are assuming
with the logit model, and how to use/interpret its results?

3) If you are OK w/ the above, take advantage of the logistic function's
symmetry (i.e., F(-B'x) = 1-F(B'x), as with the probit model) to clean
up formulas.

The only Excel-specific things I see you possibly questioning or using,
assuming you know how to start Excel, navigate its menus and windows,
and enter formulas, is the use of array formulas. Some, such as
SumProduct, are now built in to Excel as functions.

Dave Braden

wrote:
Hi,

Can anyone show me how to do the logit model overl excel? Here is the
sample data. Many many thanks!!

Flu Vaccine=1 EDUC AGE FEMALE Not Married
0 10 43 0 0
0 12 38 0 0
1 12 58 0 0
1 14 42 0 0
1 8 56 0 0
1 17 41 0 1
1 17 26 0 0
1 14 24 0 1
1 14 34 0 0
0 12 53 0 0
0 14 20 0 1
1 16 29 0 1
1 16 48 1 0
0 13 22 1 0
0 13 35 0 1
1 12 43 1 0
0 16 39 1 0
1 14 42 0 0
1 14 41 0 0


--
Please keep response(s) solely within this thread.