Hi,
I have a dataset with 5 columns: Code, Type, Site, Month, Count. I am
running bivariate regression by Code, Type, Site using the split file
option. However, I am not able to save the confidence interval for each
model. The save option saves the confidence estimates for each data point. I
also tried using the /OUTFILE = COVB ('file_name') option. This saves the
regression estimates (constant, variable value, std error etc) for each
model but not the confidence interval estimates. For the data given below
there would be 6 models for each code, type, site combination so I would
like 6 confidence interval estimates. The output gets printed on the SPSS
output screen but does not get saved to a dataset. Can anyone pls tell me
the code for the same?
Thanks
The code I used is as follows:
split file by code,type,site.
REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS CI
/CRITERIA=PIN(.05) POUT(.10) CIN(95)
/NOORIGIN
/DEPENDENT Count
/METHOD=ENTER Month
/SAVE PRED MCIN .
Sample data:
CodeType Site Month Count
1 1 1 9 12
1 1 1 10 23
1 2 3 11 19
1 2 3 12 31
2 2 4 4 11
2 2 4 5 14
2 5 1 1 15
2 5 1 2 13
2 5 8 9 20
2 5 8 11 21
3 5 9 11 33
3 5 9 12 34
Thanks!
S
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD