McKelvey-Zavoina N-Chotomous Multivariate Probit



NKOTP.ZIP contains the following 6 files:

NKOTP2.FOR -- FORTRAN program
NKOTP2.EXE -- Compiled program that will run on any Pentium computer
NKOTP.DAT -- "Control card" file that controls the operations of the program (see below)
ABORTK.DAT -- Example data file
TERM3.DAT -- Output File for example
CLASS3.DAT -- Output File for example

NKOTP2.FOR is a slightly modified version of the original McKelvey-Zavoina N-Chotomous Probit program written by Dick McKelvey in the late 1960s. The original program was written for the IBM 360 computer and used the old 130 character line fan-fold paper output. It also used direct writes to disk units (e.g., WRITE(20)....) to conserve memory (this was 30 years ago!).

These features were eliminated in this version of the program. I modified the program so it does not do unformatted writes/reads and I re-formatted the two output files so that they have standard 80 character formats.

To run the program the user must modify the "control card" file NKOTP.DAT. This file controls the program options and points the program to the data file being analyzed and what files to use for the output. Below is the example I have included with the program:

 ABORTK.DAT
 TERM3.DAT
 CLASS3.DAT
  CHECKING THINGS OUT
          RUN ZULU
   12    1    4   10  415    .00001
    0    1    2    3
 (12F13.6)
                 
0203040506070809101112
  1. The first line -- ABORTK.DAT -- is the name of the data file.
  2. The second line -- TERM3.DAT -- is the name of the file containing a copy of the output you will see on you screen when you run the program.
  3. The third line -- CLASS3.DAT -- is the name of the file containing an analysis of each observation of the data file. It has the predicted probabilities for each response category.
  4. and ...
  5. This is a two line title of your choice. Always include this even if the lines are blanks!
  6. The Numbers on the sixth line are in order:
    1. Total number of variables including the dependent variable. The maximum is currently set at 50. This can be increased. Send me E-Mail if you have a large problem you want to analyze.
    2. The position of the dependent variable
    3. The number of possible responses on the dependent variable -- dependent variable values that are not specified as one of the valid responses are treated as missing data (list-wise deletion).
    4. The number of allowable iterations
    5. The total number of observations -- Note that the maximum is 99,999. This can be increased. Send me E-Mail if you have a large problem you want to analyze.
    6. Constant whose value is used to determine the degree of precision desired.

    The format of this line must be (5I5,F10.5).

  7. Line seven lists the valid responses for the dependent variable. The format of this line must be *I5 where * is the number of valid responses.
  8. and ...
  9. This is the FORMAT specification for the input data matrix.
    Always include the second line even if the format statement uses only the first line!
  10. The order numbers of the independent variables to be included in the analysis. Note that if the Dependent variable were the 3rd column of the data matrix then these numbers would be:

    0102040506070809101112
To run the program, simply place these 6 files in a directory and type NKOTP2. The program reads NKOTP2.DAT and ABORTK.DAT and outputs TERM3.DAT and CLASS3.DAT. To run your own analysis, simply change the file names and control numbers in NKOTP.DAT.

If you need help send me E-Mail at: KPoole@ucsd.edu