home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
The Fred Fish Collection 1.5
/
ffcollection-1-5-1992-11.iso
/
ff_disks
/
300-399
/
ff386.lzh
/
Statpack
/
statpack.lzh
/
Examples
/
Framingham
/
Framingham.doc
< prev
next >
Wrap
Text File
|
1988-08-25
|
1KB
|
22 lines
The Framingham example is an already constructed frequency table from
Ku, H.H. & Kulback, S. (1974) "Loglinear models in contingency table
analysis." American Statistician 28: 115-122. The three variables are (1)
systolic blood pressure (the columns - <127, 127-146, 147-166, 167+), (2)
Serum cholesterol (the rows - <200 mg/100cc, 200-219, 220-259, >=260), and
(3) coronary heart Disease (the subtables or layers - present/absent).
Any two- or three-way frequency table such as this can be read directly
into Statpack and the log-linear model applied. We read the table as one
variable called Frequency, with 32 values and immediately apply Log Lin3.
Statpack informs us that it assumes the table to have been read row by row
for each layer, then asks for the column variable. This is the blood Pressure
with 4 categories. Next comes the row variable, Serum cholestrol with 4
categories, and finally the layer variable, coronary heart Disease with 2.
The association table of Chi-square values shows that the three-way
interaction is not significant, that the two-way effects are. We display the
no-interaction model.
If we wish to see the various classified percentages for the table, we
can now (after Log Lin3 in this case) apply Tabulate3. We use Frequency as
the weight variable and obtain the table.