ASSOC is a program for the analysis of association tables. The original code was first developed by

The following input cards are required (each card is one line)

1. (A60) Title for run

2. (10I2) Input/Output information -- Place a 1 in :

Column 2 to suppress observed, expected counts and std.

residuals

Column 4 to suppress observed odds-ratios

Column 6 to suppress expected odds-ratios

Column 8 to suppress parameter estimates

Column 10 for Model II calculations (otherwise Model I

used)

Column 12 to suppress iteration detail

Column 14 if start tables are to be input

Column 16

If no specific diagonal parameters are to be included

type in Column 16 zero. Otherwise:

1 for free diagonals, i.e. each diagonal cell in every

group has one diagonal parameter (quasi- models);

2 for Group Specific Diagonal Parameters, i.e. one

diagonal parameter for each group (table)

3 for Category Specific Diagonal Parameters, i.e. one

diagonal parameter for each diagonal cell for all

groups (same density in the same diagonal cell in each

table).

4 Combination of 2 and 3: Group and Category Specific

Effects

5 One diagonal parameter for all diagonal cells

Column 18 to suppress the printing of Row and Column

Parameters

Column 20 to suppress the printing of Alpha- and Beta-

Ratios, or Mus, Nus and Mu- and Nu-Differences.

3. (2I2,I3,F8.7,I4,F4.3) #Rows, #Cols, #Groups,

Tolerance(def=.01) Iterations (def=100), Constant to be

added to all frequencies (def=.0)

4. Input Format for data -- e.g. any valid Fortran format;

examples: (8F6.0),(8X,7F5.3), or FREE

5. Name of file on which the data are stored in the above

Format

6. Input start tables (necessary only If column 14 of card #2

is 1) -- In the same Format as was used for the data,

place a 1 in those cells to be included and a 0 (zero) in

those to be ignored or excluded, i.e. treated as

structural zeroes.

7. (15I2) Model selection : Place a 1 in :

Column 2 If Null Conditional Association (Conditional

Column 4 If Homogeneous Conditional Uniform Association

Column 6 If Heterogeneous Conditional Uniform Association

Column 8 If Homogeneous Conditional Row-Effects

Column 10 If Homogeneous Conditional Column Effects

Column 12 If Simple Heterogeneous Row Effects

Column 14 If Simple Heterogeneous Column Effects

Column 16 If Heterogeneous Row Effects

Column 18 If Heterogeneous Column Effects

Column 20 If Homogeneous Row and Column Effects

Column 22 If Heterogeneous Row, Homogeneous Column

Column 24 If Heterogeneous Column, Homogeneous Row

Column 26 If Heterogeneous Row and Column

Column 28 If Homogeneous Equal Row and Column Scalings

Column 30 If Heterogeneous Equal Row and Column Scalings

(1) We store the following data (from Clogg 1982, AJS, Table 5)

in file "abort2.dat":

44 11 38 62

59 41 147 293

23 11 13 27

27 8 16 27

258 57 105 110

(2) We edit a control file "abort2.ctl" as:

Replicating Clogg (1982), Data from Table 5

0 0 0 0 1 1 0 0 0 0

5 4 1

FREE

abort2.dat

1 0 0 0 0 0 0 0 0 0 0 0 1 0 0

In UNIX, enter:

assoc abort2.ctl abort2.dat abort2.log

Note: three arguments follow program "assoc": (1) control file,

(2) data file, and (3) output file

The following files are provided to you free of charge. You

may copy, distribute, and use these files for education and

research purposes.

assoc.f Fortran source code

abort2.ctl A test program

abort2.dat Data on abortion attitude

abort2.log Output file from abort2.ctl

abortion.ctl GLIM control file for comparison

abortion.log GLIM output file from abortion.ctl